[Pp. 111-164 in Thomas S. Weiner, ed., Discovering Successful Pathways in Children’s Development: Mixed methods in the Study of Childhood and Family Life. Chicago: University of Chicago Press, 2005.]

Sites of Belonging: Acculturation, Discrimination, and Ethnic

Identity among Children of Immigrants

Rubén G. Rumbaut

“We define our identity always in dialogue with, sometimes in struggle against, the things that significant others want to see in us… The monological ideal seriously underestimates the place of the dialogical in human life.”

--Charles Taylor (1992: 33) “ ‘Being American means that you feel like you’re the norm,’ one of my friends tells me… [But] in a splintered society, what does one assimilate to? … I want to figure out, more urgently than before, where I belong in this America that’s made up of so many subAmericas. I want, somehow, to give up the condition of being a foreigner… I have to make a shift in my innermost ways. I have to translate myself.”

--Eva Hoffman (1989:202, 210-211) “My children, who were born in this country of Cuban parents and in whom I have tried to inculcate some sort of cubanía, are American through and through. They can be ‘saved’ from their Americanness no more than my parents can be ‘saved’ from their Cubanness… Like other second-generation immigrants, they maintain a connection to their parents’ homeland, but it is a bond forged by my experiences rather than their own. For my children Cuba is an enduring, perhaps an endearing, fiction. Cuba is for them as ethereal as the smoke and as persistent as the smell from their grandfather’s cigars (which are not even Cuban but Dominican).”

--Gustavo Pérez Firmat (1994:5) “My identity is hardly clear-cut… To my parents, I am all American, and the sacrifices they made in leaving Korea…pale in comparison to the opportunities those sacrifices gave me. They do not see that I straddle two cultures, nor that I feel displaced in the only country I know. I identify with Americans, but Americans do not identify with me. I’ve never known what it’s like to belong to a community… I know more about Europe than the continent my ancestors unmistakably came from… Though they raised me as an American, my parents expect me to marry someone Korean and give them grandchildren who look like them… My parents didn’t want their daughter to be Korean, but they don’t want her fully American, either. Children of immigrants are living paradoxes. We are the first generation and the last. We are in this country for its opportunities, yet filial duty binds us. When my parents boarded the plane…I don’t think they imagined the rocks in the path of their daughter who can’t even pronounce her own name.”

--Carolyn Hwang (1998:16)

Introduction

What a difference half a century makes. Oscar Handlin’s The Uprooted, a popular portrait of immigrant America by the leading immigration historian of his time, was published 50 years ago. From his vantage at mid-century, as he wrote in a postscript to the second edition two decades later, immigration was already “a dimly remote memory, generations away, which had influenced the past but appeared unlikely to count for much in the present or future” and ethnicity, not a common word in 1950, seemed then “a fading phenomenon, a quaint part of the national heritage, but one likely to diminish steadily in practical importance” (1973:274-275). After all, the passage of restrictive national-origins laws in the 1920s, the Great Depression and World War II had combined to reduce the flow of immigrants to the United States to its lowest point since the 1820s. But history is forever being ambushed by the unexpected. As it turned out, the so-called "American century" ended much as it had begun: the United States has again become a nation of immigrants, and it has again been transformed in the process. Today, the phenomena of immigration and ethnicity are anything but fading and decreasing in importance: there are now over 30 million immigrants residing in the United States, the largest total ever, and their US-born children number around 30 million more, so that today more than 20 percent of Americans—and half of all Californians—are of “foreign stock.” They hail overwhelmingly from Asia, Latin America and the Caribbean this time, not from Europe. Never before has the United States received such diverse groups, immigrants who mirror in their origins and destinies the forces that forged a new world order in the second half of the 20th century, and who are, unevenly, engaged in the process of trying to figure out where they belong in this America that’s made up of so many sub-Americas.

The increasing size and concentration of the foreign-born population, added to its extraordinarily diverse national and socioeconomic origins and modes of incorporation, have raised significant questions about the impact of immigration on American society and led to a burgeoning research literature. Less noticed in that literature, however, or in the public debates swirling about the topic, has been the fact that a new generation of Americans raised in immigrant families has been coming of age—transforming their adoptive society even as they themselves are being transformed into the newest Americans. Over time, its members will decisively shape the character of their ethnic communities, even as the larger society that receives them will decisively shape their own ethnic identities. The long-term effects of contemporary immigration will hinge more on the pathways and trajectories of these youths than on the fate of their parents.

Yet we know surprisingly little about them (it didn’t help that the critical question on parental nativity was dropped from the last three censuses, pulling out the rug from under the study of intergenerational mobility at the very moment that a new era of mass immigration was rapidly taking hold). Consider, for example, this handful of fundamental empirical questions: Among children of immigrants, what are their patterns of English vs. non-English language use, preference and proficiency? Is there any evidence of English language assimilation over time among different national-origin groups? Is there a similar assimilative shift over time in their ethnic self-identities and sense of belonging? If so, what explains it? If not, how do they identify? Is ethnic identity mainly a function of filial-familial attachments, degree of acculturation, experiences of discrimination, or other factors?

Mixed Methods in the Contextualized Study of Ethnic Identity

In the study I will describe in this chapter, I have extracted a great deal of analytical mileage on ethnic self-identity and the factors that shape it from a single open-ended question asked at two points in

time, three years apart, in mostly structured surveys of a large representative sample of adolescent children of immigrants coming of age on both coasts of the United States. Those surveys were comprehensive in scope and supplemented by separate interviews with their parents, in which we were able to ask some equally-worded questions for comparative analysis, including items on racial self-identification. Moreover, the sampling frame on which the study is based allows us to generalize our findings to the population universes from which they were drawn, enhancing the utility of our results. Nonetheless, I remain keenly aware of the limitations involved in this analysis—both substantively and methodologically—and of the value added by the use of mixed methods in the contextualized study of ethnic identity.

Ethnic self-identities can be understood as “definitions of the situation of the self.” For children of immigrants, they emerge from the interplay of racial and ethnic labels and categories imposed by the external society and the original identifications and ancestral attachments asserted by the newcomers. They are contextually malleable and may be hypothesized to vary across different social situations, across different developmental stages throughout the life course, and across different historical contexts. Our survey instruments could not examine the situational contexts in which ethnic self-identities may be differentially deployed; yet, in theory, the way an ethnic identity may be expressed in one’s presentation of self—indeed, of multiple selves—can differ depending on one’s audience. For instance, it is possible in principle to identify as “Chicano” with one’s peers, as “Mexican American” in varying secondary relations, as “Hispanic” in filling out official forms, as “Mexican” with one’s parents at home… and ironically as “American” not in the United States but when making a visit to Mexico to see relatives who might focus on the poor Spanish or other insignia of “Americanization” that make the U.S.-based visitor suddenly conscious of cultural differences vis-à-vis their Mexican counterparts. These kinds of contextual influences on the definition and presentation of the self suggest that mixed-method work on ethnic identity should be an essential part of a comprehensive view. At the conclusion of this chapter, I return to some of our current research which involves such mixed-method work.

Furthermore, we surveyed our respondents from their mid to their late teens. But adolescence is, among other things, a developmental period of “identity crisis” and identity formation; the results reported here are thus limited to that period of the life course and may not apply to samples of, say, middle-aged adults with children of their own in whom they may be trying to inculcate a sense of ethnic attachment (an issue echoed by the quotes from Pérez Firmat and Hwang above). And as seen especially among Mexican-origin respondents with respect to reactive self-identifications expressed in the aftermath of the Proposition 187 campaign in California (discussed in what follows), the larger sociopolitical and historical context matters. But our surveys were limited to the same locales (in California and Florida) in the mid 1990s.

Methodologically, the quantified coded responses given by thousands of individuals from dozens of diverse national origins in our surveys revealed a handful of distinct patterns that advance our explanation of complex processes of self-identification. But they do not tell us much about the (likely varying) subjective meanings that particular ethnic or racial labels may have had for individual respondents. Ethnicity is likely to some degree to be racialized in the respondents’ own notions of these modes of group identity (as noted in William Cross’s chapter in this volume). For that matter, there is a strong possibility that the same label (e.g., “Latino”) may take on different meanings for the same individual at a later time (e.g., in young adulthood) or in different circumstances (e.g., in college) (cf. the chapters by Cooper and Goldenberg et al. in this volume). Such considerations underscore the need to both broaden and deepen our research by incorporating mixed methods as appropriate to get at those

dimensions of varying subjectivity and situationality, and to facilitate a more thoroughly contextualized study of ethnic identity and social belonging. In particular, mixed-method considerations suggest the need to extend our study longitudinally into adulthood, and to complement the survey methods we have employed in CILS—and the representative sample on which the study is based—with in-depth qualitative interviews, oral histories, and targeted ethnographies. In the following sections of the chapter, I turn to laying out the research design and key findings of this larger project on ethnic identity. In a final section I return to further work using mixed methods grounded in the survey studies.

The CILS Study and Sample Characteristics

The Children of Immigrants Longitudinal Study (CILS) has followed the progress of a large sample of youths representing 77 nationalities in two main areas of immigrant settlement in the United States: Southern California (San Diego) and South Florida (the Miami and Fort Lauderdale metropolitan area). The initial survey, conducted in Spring 1992 (“T1”), interviewed 5,262 students enrolled in the 8th and 9th grades in schools of the San Diego Unified School District (N=2,420), and of the Dade and Broward County Unified School Districts (N=2,635, with another 207 enrolled in private bilingual schools in the Miami area). The sample was drawn in the junior high grades, when dropping out of school is rare, to avoid the potential bias of differential dropout rates between ethnic groups at the senior high school level. Students were eligible to enter the sample if they were U.S.-born but had at least one immigrant (foreign-born) parent, or if they themselves were foreign-born and had come to the U.S. at an early age (before age ten). The resulting sample is evenly balanced between males and females, and between foreign-born (the 1.5 generation) and U.S.-born children of immigrants (the second generation) (for details see Rumbaut 1994).

Three to four years later, in 1995-96 (“T2”), a second survey of the same panel of children of immigrants was conducted—this time supplemented by in-depth interviews with a stratified sample of their parents. The purpose of this follow-up effort, which succeeded in re-interviewing 82% of the baseline sample, was to ascertain changes over time in their family situation, school achievement, educational and occupational aspirations, language use and preferences, ethnic identities, experiences and expectations of discrimination, and psychosocial adjustment. By this time the youths, who were originally interviewed in junior high when most were 14 years old, had reached the final year of senior high school (or had dropped out of school); most were 17 to 18 years old at the time of the T2 survey. [The study, still ongoing, returned to the field in 2001-03 (“T3”), a decade after the baseline survey, with the respondents now in their mid-twenties and in many cases located across the country; while results were not available at this writing, preliminary findings—this time adding in-depth, open-ended oral histories to survey methods of data collection—will be noted toward the end of the chapter.]

[Table 1 here]

As shown in Table 1, the principal nationalities represented in the San Diego CILS sample are Mexican, Filipino, Vietnamese, Laotian, Cambodian, and smaller groups of other children of immigrants from Asia (mostly Chinese, Japanese, Korean, and Indian) and Latin America. In the South Florida CILS sample, the principal national-origin groups consist of Cubans, Haitians, Jamaicans, Nicaraguans, Colombians, Dominicans, and others from Latin America and the Caribbean. The modest family origins of many of these children, the highly educated backgrounds of the parents of others, their varying patterns of employment and homeownership, are all reflected in the sample. For example, only a small proportion of Mexican and Indochinese fathers and mothers had college degrees, well below the 1990

U.S. norm of 20% for adults 25 and over. By contrast, 41% of Filipino mothers had college degrees, well above national norms. The contrast is made even sharper by the proportion of parents with less than a high school education--that is, less than what their children have now already achieved. Most of the more recently arrived foreign-born children from Mexico, Haiti, Vietnam, Laos and Cambodia had fathers and mothers who never completed secondary-level schooling. Rates of labor force participation varied widely by nationality for both fathers and mothers: the Indochinese had very low rates of labor force participation, indicative of their eligibility for and use of public assistance, whereas most of the other groups had LFP rates that exceed national norms. And there was a huge gap between ethnic groups in their proportion of homeowners, ranging in 1995 from a low of 4% among Hmong families from Laos and 8% of the Cambodians to over 80% of the Filipinos.

Table 1 provides a socioeconomic status ranking of the major ethnic groups in the CILS sample, as measured in 1992 by a standardized composite index of father’s and mother’s education, occupational prestige, and home ownership. These standardized scores are transformed to a scale of 0 to 1, with higher scores indicating higher parental socioeconomic status. As these rankings show, most advantaged were Cuban-origin students attending private schools in Miami (with a score of .612), followed by groups from Europe, Canada and the Middle East (.559) and other Asia (.551), the Filipinos (.516) and the Chinese (.501). In the middle, with SES scores hovering around the sample average of .424, were the Jamaicans and other West Indians, and all other Latin American groups except Mexico. At the bottom were the Haitians (.382) and Mexicans (.269), and all of the Southeast Asian refugee groups: the Vietnamese (.342), Lao (.233), Cambodian (.172), and Hmong (.103).

Immigrant Families and Forms of Parent-Child Acculturation

There are also significant differences in family structure among these groups, and in the character of parent-child relationships. Growing up in immigrant families is often marked by dissonant acculturation, when linguistic and other acculturative gaps develop between parents and children that can exacerbate intergenerational conflicts, cause the children to feel embarrassed rather than proud of their parents as they try to fit in with native peers, or lead to role reversals, as children assume adult roles prematurely by dint of circumstance. There are other types of intergenerational adaptation; e.g., we have also identified and distinguished between consonant and selective acculturation (Portes and Rumbaut 1996, 2001). But all immigrant families must contend not only with the “generation gaps” and the “storm and stress” of adolescence, but of acculturation to a new society as well.

[Table 2 here]

Table 2 shows several key objective and subjective indices of family structure and parent-child relations, broken down by national origin and the children’s level of acculturation and language ability. Intact family in Table 2 refers to those cases in which both biological parents were present at both the 1992 and 1995-96 surveys (61 percent). Table 2 also provides information on four subjective dimensions of the parent-child relationship: family cohesion, parent-child conflict, embarrassment over parents’ ways, and attitudes of familial obligation (for details on these scales, see Rumbaut 1994). As noted above, immigrant families and groups can be located along a continuum ranging from situations where parental authority is fully preserved to those where it is thoroughly undermined by generational gaps in acculturation—particularly in English knowledge and the extent to which second-generation youth retain their parents’ language. In operational terms these patterns should be reflected in the

degree of intergenerational cohesion or conflict between immigrant parents and their children, the extent to which these youths report being embarrassed by their parents or attached to them by filial duty.

The results in Table 2 show that Asian-origin families are comparatively more likely to remain intact (the low-SES Hmong and Cambodian refugees are the main exceptions among Asian groups in this regard, in part as a result of a greater proportion of widowed mothers than other groups). The Dominicans are the only Latin American nationality to exhibit the pattern of high family structural instability seen among the Haitians and West Indians. As Table 2 shows, there is not a significant association between family structure and the youths’ level of acculturation. But a different picture emerges when we examine the various subjective indices of family integration. Without exception, the Latin American nationalities exhibit the most cohesive families as well as the lowest levels of parent-child conflict. Most of the Latin groups also exhibit lower proportions of youths who report being embarrassed of their parents, with the lowest (14 percent) found among two groups of modest socioeconomic status, the Mexicans and the Dominicans. Indeed, a Spanish-language background was the strongest predictor of fluent bilingualism in our sample, indicative not only of the significant advantage of Spanish speakers relative to other languages in this sample, but also of its association with consonant or selective forms of acculturation among parents and children of Latin American origin. By contrast, all of the Asian, European/Canadian, and Black Caribbean groups fell below the sample average in their percent of high-cohesion families; and nearly all of the Asian and Black Caribbean groups scored above the sample average for high-conflict families. Overall, those families were more strongly associated with patterns of dissonant acculturation. However, the same rank order was not seen with the remaining index of familism. The youths who most strongly adhered to attitudes granting primacy to familial obligations were the Southeast Asians (especially the Laotians and the Vietnamese) and the Mexicans. The least familistic—i.e., those more oriented to individualistic rather than collectivistic outlooks—were the European/Canadians, most of the Latin Americans and the West Indians, and the Filipinos, Chinese and other Asians.

But as Table 2 underscores, all of the four subjective family indices were strongly associated with the youth’s level of acculturation and language dominance. The acculturation index is a composite measure of the child’s preferences for the English language and for American ways as reported at both surveys (a score of 0 indicates that no preference for either was reported at either survey, while a maximum score of 1 indicates that a preference for both was reported at both surveys). The typology of language dominance is based on the youths’ varying levels of proficiency in both English and the parental language. The data show that as the youth’s level of acculturation increases—and by implication, as acculturative gaps widen and the degree of intergenerational dissonance increases—the level of parent-child conflict and of embarrassment over parents’ ways increases, while that of family cohesion and of familistic attitudes decreases. Conversely, greater family cohesion and familism are associated with lesser acculturative preferences for English and American ways—and by implication, with consonant acculturation in intergenerational relations.

[Figure 1 about here]

These patterns are clearly graphed in Figure 1, and are reinforced by the data on language dominance. Youths classified as English dominant and limited bilinguals—that is, who are not fluent in the parental language—exhibit much higher parental conflict and embarrassment profiles, and much lower family cohesion, than do youths classified as fluent bilinguals and foreign-language dominants—that is, who are fluent in the parental language. Significantly, the familism pattern diverges here. The

fluent bilinguals and English dominants—those who are fluent in English and most acculturated—are the least familistic, while the limited bilinguals and foreign-language dominants—those who are neither fluent in English nor highly acculturated—are much more likely to adhere to familistic attitudes.

Language Maintenance and Language Shift

Over 90% of these children of immigrants reported speaking a language other than English at home, mostly with their parents. But at the T1 survey in 1992 already 73% of the total sample preferred to speak English instead of their parents’ native tongue, including 64% of the foreign-born youth and 81% of the U.S.-born. By the T2 survey three years later, the proportion who preferred English had swelled to 88%, including 83% of the foreign-born and 93% of the U.S.-born Even among the Mexican-origin youth living in San Diego—a city on the Mexican border with a large Spanish-speaking immigrant population and a wide range of Spanish-language radio and TV stations—the force of linguistic assimilation was incontrovertible: while at T1 only a third (32%) of the Mexico-born children preferred English, by T2 that preference had doubled to 61%; and while just over half (53%) of the U.S.-born Mexican-Americans in San Diego preferred English at T1, that proportion had jumped to four-fifths (79%) three years later. More decisively still, among Cuban-origin youth in Miami, 95% of both the foreign-born and the native-born preferred English by T2.

A main reason for this rapid language shift in use and preference has to do with their increasing fluency in English (both spoken and written) relative to their level of fluency in the mother tongue. Respondents were asked to evaluate their ability to speak, understand, read and write in both English and the non-English mother tongue; the response format (identical to the item used in the U.S. census) ranged from “not at all” and “not well” to “well” and “very well.” Over three-fourths of the total sample at both T1 and T2 reported speaking English “very well,” compared to only about a third who reported an equivalent level of spoken fluency in the non-English language. Even among the foreign born, those who spoke English very well surpassed by 69% to 41% those who spoke the foreign language just as well. And the differences in reading fluency were greater still: those who could read English “very well” tripled the proportion of those who could read a non-English language very well (78% to 24%).

As a consequence, the bilingualism of these children of immigrants becomes increasingly uneven and unstable. The CILS data underscore the rapidity with which English triumphs and foreign languages atrophy in the United States, as the second generation not only comes to speak, read and write it fluently, but prefers it overwhelmingly over their parents’ native tongue. It bears adding that these results have occurred while the youths were still residing as dependents in their parents’ home, where the non-English mother tongue retains primacy; once they leave the parental fold to lead independent lives of their own, the degree of English language dominance and non-English language atrophy is almost certain to accelerate, all the more among those living outside dense immigrant enclaves.

But what about their sense of identity, of belonging and ethnic loyalty? Does it too shift rapidly, like language, in straight-line fashion toward dominant American models as these youths seek to fit in with their new social settings? In the jargon of sociology, what is their “definition of the situation”? Perhaps the best way to begin this analysis is with a story.

The Question of Identity

In Los Angeles a few years ago, a well-acculturated Southern California native and high school senior, Stephanie Bernal, thought of herself as a “mixed chocolate swirl.” Half Latin and half Anglo by her reckoning, she had not had more than a fleeting connection with Mexico, her mother’s homeland, until Proposition 187 exacerbated ethnic tensions in California in the fall of 1994. The measure aimed to “Save Our State,” as Proposition 187 was called, by denying social and non-emergency health care services—and access to public schools—to undocumented immigrants and their children. It also required school districts to verify the legal status of students’ parents or guardians and to report to state officials any persons suspected of being in the United States unlawfully so that they may be detained and deported. Stephanie reacted by joining with friends who were organizing her school’s anti-187 movement and by affirming, even reveling in, the identity of her maternal ancestry: “When we get together to talk about it, we speak Spanish and just feel good about being Mexican” (Pyle and Romero 1994).

The week before, more than 70,000 people had marched in protest through downtown Los Angeles—the largest such march in memory, which was heavily covered on television and by the news media—spawning subsequent marches and student walkouts in local campuses. To the marchers, who included many second-generation youths from area high schools, Proposition 187 was an affront to their parents, their friends and neighbors, according to local news accounts of the events at the time (McDonnell 1994; Navarette 1994; Pyle and Romero 1994). Parents too joined in some of the marches or encouraged their children’s participation, making the initiative an issue through which the budding teenage activists could bridge the generational divide and express themselves in solidarity with their parents.

Proposition 187 was attacked by activists as a measure which would invite if not incite pejorative stereotyping and discrimination against anyone fitting a “brown” racial profile. The initiative was being advanced in a context of rising anxiety about the changing racial-ethnic composition of California, which, it was regularly noted, was expected to become the nation’s first “majority minority” state shortly after the year 2000. Referring to “that commercial,” a frequently repeated ad on television which showed people running across the California-Mexico border, sophomore Jorge Higareda complained that “they show Mexicans, but they don’t show Asians coming over in boats or anybody else. It’s like we’re the only ones coming here. And then they call us illegal.” Students at another high school insisted that “this is mostly a Latino school and everyone’s proud of their heritage—they want to defend it.” In yet another over 200 students walked out of school in protest; and, as in the huge march that had preceded it, the teenagers carried banners saying “No on 187”—and waved Mexican flags (Pyle and Romero 1994).

The red, white and green flags were meant as a spontaneous show of pride and self-validation by people who felt that their Mexican heritage was under siege; but to non-Latino voters the flags were seen as symbols of anti-American defiance and of suspicious allegiance to a foreign country. A former Marine, watching the high school marchers, reacted with anger: “You see these kids walking down the street with a Mexican flag and demanding that this country give them a free education. It aggravates people.” But a 15-year-old sophomore, Mariela Flores, shot back: “Why should we carry American flags when this country wants to kick my friends out of school and send them back to Mexico? We carry Mexican flags because they give us pride in where we come from” (Banks 1994).

Proposition 187 won in a landslide, getting 59% of the statewide vote; in populous San Diego and Orange Counties, south of Los Angeles, the measure passed with 67% of the votes cast. But Stephanie’s Mexican ethnic self-identity was “thickened” in the process, a sense of belonging made more salient than ever as she came to define who she was and where she came from in opposition to who and what she was not. The divisive campaign had the unintended consequences of accentuating group differences, heightening group consciousness of those differences, hardening ethnic identity boundaries between “us” and “them,” and promoting ethnic group solidarity and political mobilization.

Modes of Ethnic Identity Formation

This process of forging a reactive ethnicity in the face of perceived threats, persecution, discrimination and exclusion is not uncommon. On the contrary, it is one mode of ethnic identity formation, highlighting the role of a hostile context of reception in accounting for the rise rather than the erosion of ethnicity (see Aleinikoff and Rumbaut 1998; Portes and Rumbaut 1996). A few years earlier, for example, second-generation Korean-Americans saw over 2,300 Korean-owned stores in Los Angeles’ Koreatown targeted by African Americans and burned during the rioting that followed a “not guilty” verdict in the 1992 trial of four white police officers charged with the brutal beating of a black motorist. The event caused many young Koreans born or raised in the United States to become self-conscious about their common fate and distinctiveness as Koreans. They reacted by participating in multigenerational solidarity rallies and by moving to organize politically to protect the interests of the parent generation and the image of the group in the larger society (Min 1995; Bozorgmehr, Sabagh, and Light 1996).

In contrast, conventional accounts of ethnic identity shifts among the descendants of European immigrants, conceived as part of a larger, linear process of assimilation, have pointed to the “thinning” of their ethnic self-identities in the United States. For their descendants, at least, one outcome of widespread acculturation, social mobility and intermarriage with the native population is that ethnic identity became an optional, leisure-time form of “symbolic” ethnicity (Alba 1990; Gans 1979; Waters 1990). As the boundaries of those identities become fuzzier and less salient, less relevant to everyday social life, the sense of belonging and connection to an ancestral past faded “into the twilight of ethnicity” (Alba 1985).

This mode of ethnic identity formation, however, was never solely a simple linear function of socioeconomic status and the degree of acculturation—that is, of the development of linguistic and other cultural similarities with the dominant group—but hinged also on the context of reception and the degree of discrimination experienced by the subordinate group. Milton Gordon, in his seven-stage portrayal of assimilation in American life, saw “identificational assimilation”—a self-definition as an unhyphenated American—as the culmination of a complex sequence made possible only if and when it was accompanied by an absence of prejudice and discrimination in the core society (Gordon 1964; Warner and Srole 1945). Whether ethnicity will become similarly optional—a matter of individual choice—for the descendants of immigrants who are today variously classified as non-white, or whether they will be collectively channeled into enduring, engulfing, racially marked subordinate statuses and forge oppositional identities, remain open empirical questions.

Drawing on the European experience and on the eve of the new immigration from Asia and Latin America, the prevailing view of the matter was framed succinctly by Nahirny and Fishman: “the erosion of ethnicity and ethnic identity experienced by most (but not all) American ethnic groups takes place in

the course of three generations… ethnic heritage, including the ethnic mother tongue, usually ceases to play any viable role in the life of the third generation” (Nahirny and Fishman 1996: 266). However, compared to language loyalty and language shift, generational shifts in ethnic self-identification are far more conflictual and complex. To those authors the “murky concept of ethnic identification” did not lend itself to intergenerational analysis along a unidimensional attitudinal continuum since “fathers, sons, and grandsons may differ among themselves not only in the degree but also in the nature of their identification with ethnicity” (Nahirny and Fishman 1996: 267).

Thus, paradoxically, despite rapid acculturation, as reflected in the abandonment of the parental language and other ethnic patterns of behavior, the second generation remained more conscious of their ethnic identity than were their immigrant parents. The parents’ ethnic identity was “so much taken for granted and accepted implicitly” that they were scarcely explicitly aware of it, but the marginality of their children made them acutely self-conscious and sensitive to their ethnicity, “especially when passing through adolescence.” Moreover, at least under reigning conditions of dissonant acculturation, “the generational discontinuity between the formative experiences and dominant environments of most immigrant fathers and sons rendered the family ineffective as an agency for the transmission of traditional ethnicity. So pronounced was this generational gap that by the time the sons reached adolescence the immigrant family had become transformed into two linguistic sub-groups segregated along generational lines.” Finally, by the third generation “the grandsons became literally outsiders to their ancestral heritage,” and their ethnic past an object of symbolic curiosity more than anything else. “There was no doubt about the national identity of the grandsons—they were simply Americans of one particular (if not mixed) ethnic ancestry. Neither was there any trace left of the ‘wounded identity’ of the sons, for in contrast with the sons, the grandsons had never experienced the full brunt of marginality” (Nahirny and Fishman 1996: 277-278).

Complex Allegiances and Contrasting Identities

Indeed, as is the case with respect to language maintenance and language shift, the decisive turning point for change in ethnic and national self-identities can be expected to take place in the second, not in the first generation. For the children, the process of “becoming American” today has itself taken a new turn, and may now include the adoption or rejection of such officially constructed pan-ethnic categories as “Hispanic” and “Asian/Pacific Islander,” which lump together scores of nationalities into one-size-fits-all minority group labels. Relative to the first generation, the process of ethnic self-identification of foreign-born immigrant children and the U.S.-born children of immigrants is more complex, and often entails the juggling of competing allegiances and attachments. Situated within two cultural worlds, they must define themselves in relation to multiple reference groups (sometimes in two countries and in two languages) and to the classifications into which they are placed by their native peers, the schools, the ethnic community, and the larger society.

Pressure from peers and from parents can tauten the tug-of-war of ethnic and national loyalties, contributing unwittingly to a sense of marginality. This state of affairs can be further complicated when identities are racialized, group boundaries are sharpened by a visible color line, and the metaphorical definition of who one is on the “inside” conflicts with the definition of who one is on the “outside,” raising questions about the “authenticity” of either identity and doubts about fully belonging to any group. Thus, for example, a young Korean woman feels the sting of being called a “twinkie” (“yellow on the outside, white on the inside”) by co-ethnics during her freshman year in college “just because I grew up in a white suburb and was a cheerleader,” while her mother does not let her forget that she has

to marry a Korean (she had a dream in English that she was “dating a non-Korean guy and my dad showed up speaking Korean”). More complicated is the identity juggling act of 17-year-old José Mendoza, a U.S.-born, Spanish-speaking Dominican who is not black enough for many African Americans, not light enough for most Hispanics, and advised by parents to “marry light.” The way he figures it, “From the inside we’re Dominicans; from the outside we’re black.” And while with his friends he talks mostly in “black Spanglish”—a mix of Spanish and hip-hop English—he feels that “I’m still part Dominican. That’s my nationality. If you become African American you give your nationality away. That’s like saying you’re betraying your country” (Veerasarn 1999; Escobar 1999).

To their parents, who came of age before they came to America, such concerns are almost entirely irrelevant. Immigrants who arrive as adults seldom lose their original linguistic allegiances or accents, even while learning English; neither, for that matter, do they readily shed their homeland memories and self-images, even after becoming naturalized American citizens. On the contrary, they come with pre-existing and fully-formed identities, along with their hopes for the future. Their experience of migration, nostalgia for the homeland, desire to instill in their children a sense of pride in their cultural heritage, and anxiety over their children’s rapid Americanization may all actually deepen the parents’ own sense of identification with “home,” perhaps most poignantly among exiles and sojourners who migrate reluctantly and sustain ritually affirmed hopes of an eventual return. The central theoretical and empirical question is rather what happens to their children and how it is that they come to define their ethnic identities and sites of belonging, particularly during their passages to adulthood—the youthful years of “identity crisis” and heightened self-consciousness when the self-concept is most malleable (Erikson 1968; Rosenberg 1965; Phinney 1990). That developmental process can be complicated for U.S.-reared children of immigrants by experiences of intense acculturative and generational conflicts as they strive to adapt in American contexts that may be racially and culturally dissonant, and in family contexts where the differential acculturation of parents and children may take a variety of forms.

Youths see and compare themselves in relation to those around them, based on their social similarity or dissimilarity with the reference groups that most directly affect their experiences—especially with regard to such socially visible and categorized markers as gender, phenotype, accent, language, name and nationality (cf. Tajfel 1981). Their social identities, forged in terms of those contrasts with others, represent the way they self-consciously define the situation in which they find themselves and construct an ongoing account of who “we”—and “they”—are. Ethnic identification begins with the application of a label to oneself in a cognitive process of self-categorization, involving not only a claim to membership in a group or category, but also a contrast of one’s group or category with other groups or categories. Such self-definitions also carry affective meaning, implying a psychological bond with others that tends to serve psychologically protective functions—indeed, they may be adopted in part because they protect the individual’s self-esteem, whereas negative self-esteem can precipitate an alteration of identification (cf. Deaux 1996). Ethnic self-awareness is heightened or blurred, respectively, depending on the degree of dissonance or consonance of the social contexts which are basic to identity formation. For majority-group youths in an ethnically consonant context, ethnic self-identity tends to be taken for granted and is not salient; but contextual dissonance heightens the salience of ethnicity and of ethnic group boundaries (cf. Rosenberg 1979). People whose ethnic, racial or other social markers place them in a minority status in their group or community are more likely to be self-conscious of those characteristics. Youths may cope with the psychological pressure produced by such dissonance by seeking to reduce conflict and to assimilate within the relevant social context—the modal response of the children of European immigrants in the American experience. An alternative

reaction may lead in an opposite direction to the rise and reaffirmation of ethnic solidarity and self-consciousness, as exemplified by Stephanie Bernal in the vignette sketched earlier. For Stephanie, the newfound sense of belonging she experienced in joining the movement against Proposition 187 helped dissolve the marginality and ambiguity of a “mixed chocolate swirl” identity.

In what follows, I examine various factors shaping ethnic self-identification in our sample of immigrant-origin youths in late adolescence. The way that these youths define themselves is significant, revealing much about their social attachments (and detachments) as well as how and where they perceive themselves to “fit” in the society of which they are its newest members. Self-identities and ethnic loyalties can often influence long-term patterns of behavior and outlook as well as inter-group relations, with potential long-term political implications. Accordingly, I hypothesize that the process of language learning and acculturation, as well as experiences of racial-ethnic discrimination, among other factors, will be accompanied by changes in the character and salience of ethnicity—ranging from "linear" to "reactive" forms, from “thick” to “thin” identities—and hence by divergent modes of ethnic self-identification. In particular, I address these questions: What are their ethnic (or pan-ethnic) self-identities? How salient or important are they to these youths, and how have they shifted over time? What characteristics distinguish the different types of ethnic (or pan-ethnic) identities from each other? And among those characteristics, which are the main predictors of different types of ethnic (and pan-ethnic) identities?

Forms of Ethnic Self-Identification

In both the 1992 and 1995-96 surveys, an open-ended question was asked to ascertain the respondent’s ethnic self-identity. No closed categories or checklists were provided, requiring the respondents to write their answers in their own words and in their own hand. Those written self-designations were then coded and quantified, and they comprise the main outcomes that I seek here to describe and explain. From the variety of responses given, four mutually exclusive types of ethnic self-identities became apparent, which accounted for over 95% of the answers given in both surveys. These four types were classified as follows: (1) a foreign national identity (e.g., Mexican, Vietnamese); (2) a hyphenated-American identity, explicitly recognizing a single foreign national origin (e.g., Cuban-American, Filipino-American); (3) a plain American national identity, without a hyphen; and (4) a pan-ethnic minority group identity (e.g., Hispanic, Latino, Chicano, Black, Asian). The first two of these identify with the immigrant experience and original homeland, if at different degrees of closeness; whereas the last two types are exclusively identities “made in the USA.” The first three also involve chiefly national identifications (past or present, or a bridging of both); the fourth reflects a denationalized identification with racial-ethnic minorities in the U.S. and self-conscious differences in relation to the white Anglo majority population.

Ethnic Identity Shifts

Figure 2 presents the frequency distribution of the results from both surveys for the longitudinal sample as a whole. In the 1992 survey, when these youths were in the 8th and 9th grades, just over a quarter (27%) identified by the foreign national origin; a plurality (41%) chose a hyphenated-American identification; 13% identified as American; and 16% selected pan-ethnic or racialized self-identifications. By the 1995-96 survey, as they were finishing high school, over a third (35%) of the same respondents now identified by national or ethnic origin; less than a third (31%) chose a hyphenated-American identification; only 3.5% identified as American; and over a quarter (27%)

selected pan-ethnic self-identifications. Figure 3 breaks down the results for the specific types of pan-ethnic identities reported at both surveys; each type except “Chicano” (which decreased from 2.2% to 0.8%) reflected increases during this period. “Hispanic” was reported by 12% of the sample in the first survey and by 18% in the follow-up; a “Latino” self-designation was selected by a much smaller number of youths of Latin American origin, but it increased more rapidly from 0.5% to 1.4%. While hardly any youth had identified as “Asian” or “Asian American” in the first survey (0.3%), that category grew noticeably to 4.5% in the second survey. “Black” self-identities doubled from 0.9% to 1.9%.

[Figures 2 and 3 here]

These results show substantial change over time, underscoring the malleable character of ethnicity. The magnitude of the change, however, is moderate: nearly half (44%) of the youths reported exactly the same ethnic self-identity in their written responses to the open-ended question in the follow-up survey. This degree of stability takes on added significance when it is considered that several years had passed between surveys, spanning a relatively volatile developmental period from middle to late adolescence in changing cultural contexts. If ethnic identity does not emerge here as a “fixed” characteristic, then, neither is it so “fluid” as to fit what E. L. Doctorow’s character, Billy Bathgate, called his “license-plate theory of identification,” the idea that “maybe all identification is temporary because you went through a life of changing situations” (quoted in Deaux 1996: 792). Still, the fact that over half of the respondents reported a change in their ethnic self-definitions in the span between surveys underscores the need not to take such identities as givens, but to track their evolution across the life course.

Moreover, the direction of the shift was unexpected. If their rapid language shift to English documented previously were to have been accompanied by a similar acculturative shift in ethnic identity, then we should have seen an increase over time in the proportion of youths identifying as American, with or without a hyphen, and a decrease in the proportion retaining an attachment to a foreign national identity. But as shown in Figure 2, the results of the 1995 survey (conducted in the months after the passage of Proposition 187 in California) pointed in exactly the opposite direction. In 1992, over 53% identified as American or hyphenated-American, but only 34% did so three years later—a net loss of nearly twenty percentage points. Meanwhile, both the foreign national origin and pan-ethnic identifications combined for a net gain of almost twenty percentage points. The shift, therefore, has been not toward mainstream identities but toward a reaffirmation of the immigrant identity for some groups (notably Mexicans and Filipinos in California, Haitians and Nicaraguans in Florida), and toward pan-ethnic minority group identities for most others, who become increasingly aware of and adopt the ethno-racial markers in which they are persistently classified by the schools and other American institutions. How and why this occurred is a complex story. Part of it has to do with the stability and salience of different types of ethnic self-identities, and I turn first to that issue.

[Table 3 here]

Stability and Salience

Table 3 presents a detailed cross-tabulation of these results. For each of the identity types chosen in 1992, the table shows the proportion of respondents who reported exactly the same identity in 1995-96 (a measure of the stability of that identity), as well as the proportions who shifted to a different type of ethnic self-identification. As Table 3 shows, 58% of the youths who identified by their own or

their parents’ foreign national origin in 1992 did so again three years later (the most stable and “fixed” type of self-identification), compared to 47% of those choosing pan-ethnic identities, 45% of those reporting a hyphenated-American identity, and a mere 15% of those who had identified as plain Americans (by far the most unstable and “fluid” type of self-identification). Of those shifting to other identities, Table 3 suggests the patterning of the shifts: they either added a hyphen or dropped the hyphen in favor of a pan-ethnic or a foreign nationality identity. Very few shifted to a plain American identity by the second survey (overall it plummeted from 13% to 3.5%); among those who had initially identified as American only, a third shifted to a hyphenated-American identity, while another third adopted a pan-ethnic type.

[Figure 4 here]

In the follow-up survey, the youths were asked how important their reported ethnic self-identity was to them. The responses, on a 3-point scale from “very important” to “not important,” provide a measure of the salience of the different identity types. Figure 4 compares the different identity types by their degree of stability (the percent reporting the same identity years later) and salience (the percent reporting their identity as “very important”). Once again foreign national-origin identities command the strongest level of allegiance and attachment: over 71% of the youths so identifying considered that identity to be “very important” to them, followed by 57% of the hyphenates, 53% of the pan-ethnics, and only 42% of those identifying as plain American. The latter emerges as the “thinnest” identity, with the lowest stability and salience scores—fitting theoretical expectations for a highly acculturated, majority-group self-image in socially consonant contexts. The foreign national identity, in contrast, emerges as the “thickest” identity, with the highest stability and salience scores—fitting theoretical expectations for a less-acculturated, more acutely self-conscious image in socially dissonant contexts. In the salience-stability hierarchy of ethnic identities, the pan-ethnic and hyphenated-American types fall in between those two poles.

Multivariate logistic regressions (not shown) were carried out to identify the factors most likely to predict ethnic identity and salience. A large array of possible predictors was examined—but only a handful of variables emerged as significant determinants of stability and salience. Identities tended to be most stable among less-acculturated respondents (i.e., who reported lesser preferences for English and American ways of doing things) and those high in perceptions of discrimination (i.e., who reported experiences and expectations of unfair treatment because of one’s ethnicity or race). Identities were also more stable among those youths whose parents were both born in the same country, and who spoke the parental language at home. Females were also more likely than males to have retained the same ethnic identity over time. Those same predictors also determined salience, except that gender had no significant effect on salience whereas our measure of family cohesion did—i.e., net of other factors, the more cohesive the family, the more likely were the youths to consider their ethnicity as being very important to them. Put another way, “thicker” (stable and salient) ethnic self-identities were least likely to be found among second-generation youth who experienced little discrimination, had become more acculturated and spoke English in homes with low family cohesion—a recipe for dissonant acculturation and ancestral dis-identification.

Ethnic Self-Identities by National Origin

How did these patterns of ethnic self-identification vary among the major nationalities in the study? Table 4 provides a breakdown by national origin, showing both the percent selecting the four

main ethnic identity types in the most recent survey, as well as the percent change (+/-) in each identity since the initial survey. In addition, “internal ethnicity” (see Bozorgmehr 1997) is specifically considered in this table by separating the Vietnamese and the ethnic Chinese from Vietnam, and the Lao and the Hmong from Laos—the ethnic Chinese and the Hmong were segregated and disparaged minority groups in their country of origin, and as such their identity choices and homeland attachments bear closer comparative scrutiny. A number of points merit highlighting.

[Table 4 here]

First, a glance at Table 4 reveals that the patterns discussed so far do not apply equally to every nationality, and that in fact there are very large discrepancies in some cases. For example, I found that 35% of the total sample identified by a foreign national origin in the latest survey, a gain of 7% from the first survey. But that figure is an average that ranges from a high of 67% among the Lao (compared to 48% of the Hmong), 58% of the Vietnamese (compared to 26% of the ethnic Chinese-Vietnamese), and over 50% of the Filipinos and Nicaraguans, to a low of only 5% of the Dominicans and 6% of the Cubans in private schools. The Mexicans and Filipinos—in California—registered the strongest gains in foreign national-origin identities from the baseline survey (each increasing by 24%), vividly documenting (especially in the Mexican case) the Prop-187-induced process of reaction formation sketched in my earlier vignette. Most of the Mexican shift came from youths who had identified as Chicano, Latino or Hispanic in the 1992 survey.

Other groups registered significant losses in national-origin identification, especially Dominicans, West Indians, and the Hmong, Chinese and smaller Asian-origin groups. All of these smaller groups instead posted very large increases in pan-ethnic identities, with 72% of the Dominicans choosing “Hispanic” or “Latino” self-identities, as did 62% of the Colombians. The Hmong and the Chinese, virtually none of whom had identified pan-ethnically as “Asian” in 1992, made large inroads into this type of self-identification by 1995. This result is clearly not explained by socioeconomic factors, since the Hmong had the lowest family socioeconomic status of all the groups in this study (and the highest poverty rate in the United States, for that matter), whereas the Chinese in this sample mostly came from professional families. By contrast, virtually none of the Filipinos identified in pan-ethnic terms (less than 2%), as did only 4% of the Haitians. The Filipinos and Haitians stand out from all other groups in that they almost entirely identified either by national-origin or as hyphenated-Americans, maintaining an explicit symbolic attachment to their parents’ homeland.

Every group, with one exception, posted losses in the plain American identity by the latest survey. Even upper-middle-class private-school Cubans, over a third of whom had identified as American in 1992, abandoned that identity almost entirely by 1995-96 (a loss of 33%). The sole—and telling—exception involve the Europeans and Canadians: not only were they the only groups who increased (by 16%) their proportion identifying as plain American, but by the last survey over half (58%) had adopted that majoritarian, mainstream identity as their own. All other groups were in the low single digits by comparison, marking a sharp segmentation of identities. Socioeconomic status alone would not explain this divergence; as noted, the high-status private-school Cubans dropped that identity, and another relatively high-status group—the Filipinos—consistently posted miniscule numbers identifying as unhyphenated Americans. Neither would acculturation to American ways by itself explain it; again the Filipinos, easily the most highly acculturated Asian-origin group, serves as a counter-example. In the following I attempt to disentangle this puzzle.

Nativity, Family, and Identity

Ethnic identity is, in part, a way of answering the question “where do I come from?” The answers given are often expressed in a metaphorical language of kinship (e.g., “homeland,” “fatherland,” “mother tongue,” “blood ties”) with reference to a “birth connection” (cf. Cornell and Hartmann 1998:54; Horowitz 1985) to nation and family—to an imagined common origin or ancestry. Even the “thinnest” ethnicities tend to be rooted in such kinship metaphors. Thus, while ethnic identities may be socially and politically constructed, they are experienced and expressed as “natural.” In this regard, nativity variables (where one was born, to whom one was born, where one’s parents were born) are clearly important to ethnic and national self-definitions. For children of immigrants, they are also variables that can significantly complicate a clear-cut answer to basic questions of ethnic self-definition, particularly when the parents’ country of birth differs from that of the child and (in cases of interethnic marriage) from each other. The extent of such differences in the nativity patterns of our respondents and their parents is depicted in Table 5, broken down by national origin groups.

[Table 5 here]

Half of the children of immigrants in the CILS sample were born in the United States (the second generation), while the other half was foreign-born (the 1.5 generation). However, as Table 5 shows, there are wide differences by national origin. Well over 90% of the Laotians, Cambodians and Nicaraguans were foreign-born, as were 84% of the Vietnamese. By contrast, over 90% of the Cubans in private school were U.S.-born, as were 86% of the Europeans and Canadians, and well over half of the Mexicans, Filipinos, and other Asians and West Indians. Note that only three-fourths of the children in the sample had parents who were both born in the same country—and hence who could transmit to their children a common national origin. These rates of homogamy vary widely by nationality, with the highest proportions of endogamous co-national parents (between 85% and 95%) found among the Laotians, Vietnamese, Haitians, Nicaraguans, and Cubans in private school, while less than one fourth (23%) of the Europeans and Canadians had co-national parents—indicative of high levels of interethnic marriage.

[Figure 5 about here]

How do these nativity patterns affect ethnic self-identification? That question is vividly answered in Figure 5, which graphs each of the main types of ethnic self-identity by the place of birth of the child and his/her parents. Among those who reported a foreign national-origin identity, 75% were foreign-born, compared to 31% of those who reported a hyphenated-American identity and a miniscule 8% of those identifying as plain American. Less than half (47%) of those identifying by pan-ethnic categories were foreign-born. In families where both parents came from the same country of birth, their offspring were much more likely to incorporate that national origin as part of their own identity (either wholly or as a hyphenated-American identity). In families where both parents were foreign-born but came from different countries of birth, their offspring were more likely to simplify the complexity of mixed origins by adopting either a pan-ethnic (minority) or plain American (majority) identity, thus resolving the conflict of identifying with one parent rather than the other. For example, asked how she identified, Rosa, the 15-year-old U.S.-born bilingual daughter of a Cuban-born father and a Salvadoran-born mother told me, “If I said I’m Cuban, that’d make my mom mad; and if I said I’m Salvadoran, my dad would get all upset, so I think I’m… Hispanic!” The pan-ethnic identity label “made sense” for her as a simplified category that focused on shared traits (and was popular with her Spanish-speaking school

friends as well) while avoiding the dilemma of privileging one or the other parental “birth connection”—to have done so would have been tantamount, to this offspring of a hybrid marriage, to an act of ethnic disloyalty to one or the other parent.

Language, Acculturation, and Identity

Of course, even when the nativity patterns of the youths and of their parents are held constant, many other factors are likely to impinge independently on the process of ethnic self-identification. Table 6 lists a variety of factors related to patterns of acculturation of both parents and children (measured in the 1992 survey) that are strongly predictive of the children’s identity outcomes (measured three to four years later in the 1995-96 survey). A glance at the data in this table makes clear that each identity type has a distinct social profile that differs significantly from each of the others. For example, the parents’ length of residence in the United States provides an indicator of their own level of acculturation. As Table 6 shows, the longer the parents had resided in the United States, the more likely their children were to assimilate their identities along the continuum from foreign national to pan-ethnic to hyphenated-American to plain American. That same linear progression is evident for each of the different indicators of parental socioeconomic status shown (home ownership, occupation and education), of foreign and English language proficiency and use, and of preference for “American ways” reported separately by both the youths and their parents. Language, in particular, is closely connected to the formation and maintenance of ethnic identity. The shift to English is associated with a shift in self-definition and seems to entail abandoning not only a mother tongue but also a personal identity. Those youths most loyal to the mother tongue were also most loyal to a national-origin identity; conversely, those youths who preferred English and spoke only English with their parents and close friends were much more likely to identify as an unhyphenated American.

[Table 6 about here]

Finally, Table 6 presents selected indicators pertaining to the youths’ attitudes toward and attachment to their parents and family, and the effect of those on ethnic self-identifications. One item asked at both surveys—measuring the youths’ response to a question about feeling embarrassed about their parents’ ignorance of American ways—provides an empirical indicator of dissonant acculturation within the family. As Table 6 shows, again the same linear relationship obtains as with the other measures of acculturation: least embarrassed were youths who identified by national origin, and most embarrassed were those who identified as plain American (with the proportion widening over time). A different set of items tapped attitudes expressing felt collective obligations toward the family; two of them are shown in Table 6 (believes that only relatives can help with serious problems, and that living close to family is very important). The higher the level of agreement on these items, the more likely the youth identified by national origin; the less the agreement, the more likely the youth identified as plain American, with the other two identity types falling in between.

But perhaps most revealing of the effect of parental ethnic socialization and of the strength of parent-child bonds on the children’s ethnic self-definitions were two open-ended questions (asked in the 1992 survey) asking the respondent to write in how they perceived their father and mother to self-identify. Those responses were coded and quantified, and correlated against the children’s own ethnic self-identities reported in the 1992 and 1995-96 surveys. Table 7 presents the correlations for the main identity types over time, which effectively provide a measure of the degree of consonant or dissonant identification of the youths with their parents. It is remarkable to see that even after the passage of

several years, and despite the attenuation of the correlation coefficients over this time, the way the children had perceived their parents’ ethnic identities remained significantly associated with their own modes of self-identification, mirroring them as if they were reflections in an ethnic looking-glass. The strongest consonant identification (i.e., the strongest correlation) involved national-origin identities, which, as seen above, were also the most stable and salient identity type.

[Table 7 about here]

Clearly the family plays a central role in the process of ethnic socialization. It is the crucible of the child’s first notions of belonging and of “home,” and it is within the family that the child first learns about and forms ethnic attachments and self-concepts. Even in the sometimes bumpy passage through adolescence and under varying degrees of differential acculturation, how these youths think and feel about themselves continue to be affected by the parents' modes of ethnic socialization and by the strength of the attachment that the child feels to the parents.

Discrimination, Reactive Ethnicity, and Identity

The story of the forging of an ethnic self-identity in the second generation, however, plays out on a much larger stage than that of the family. I began by focusing attention on social forces outside the family that shape both the creation of racial-ethnic categories and of ethnic self-definitions, particularly those involving discrimination and the politics of reactive ethnicity and not only those involving acculturation and the psychology of linear ethnicity. I now return to those concerns.

Regional location (in Southern California or South Florida) and the type of school these youths attended (inner-city or suburban, public or private schools) are two such extra-familial contextual factors. They delimit the youths’ exposure to different social worlds, shape differential associations with peers in those contexts, and influence attendant modes of ethnic socialization and self-definition. As Table 8 shows, slightly over half of the sample (52%) was located in South Florida, and slightly less than half (48%) in the San Diego area. But two-thirds (67%) of those identifying by a foreign national origin were in San Diego, while about four-fifths (79%) of those identifying as plain American were in South Florida. The distribution of the main types of pan-ethnic identities—as specified in Table 8—is even more skewed by location, reflecting the varieties of groups that have settled in those areas. The location of the school attended in junior high school—which is in part a function of family socioeconomic status—remains associated with the self-identities reported by the youths several years later. Those identifying as plain Americans were most likely to have attended suburban schools (73%), followed by those identifying as hyphenated-Americans (70%); by comparison, the youths identifying by each of the different varieties of pan-ethnic labels were more likely to have attended inner-city schools with peers who tend to be primarily native racial and ethnic minorities. Although only a third (35%) of the sample attended schools in the inner city, half of those identifying as Black did so (49%), as did 43% of the Asians, and 38% of the Hispanics or Latinos.

[Table 8 about here]

The process of growing ethnic awareness among the children of immigrants in the CILS sample is especially evident in their perceptions, experiences and expectations of racial and ethnic discrimination. As Table 8 shows, reports of actually being discriminated against increased from 54% to 62% of the sample in the last survey. Virtually every group reported more such experiences of

rejection or unfair treatment against themselves as they grew older, with the highest proportions found among the children of Afro-Caribbean and Asian-origin immigrants, followed by Mexicans and other Latin Americans, and the lowest proportions among Cuban youth in Miami.

As seen in Table 8, there are significant associations between experiences and expectations of discrimination reported at both surveys, and the types of ethnic and pan-ethnic self-identities expressed in the latest survey. Indeed, discrimination sharpens ethnic-racial identity boundaries and increases the salience of the category on the basis of which persons experience unfair treatment. Among the three main types of ethno-national identities, the highest proportions of experienced and of expected discrimination were reported by those who identified by national origin, and the lowest by those who identified as American, with hyphenates in between. Among the three main types of pan-ethnic identities, by far the highest proportions were seen among “Black”-identified respondents, with “Asian”-identified pan-ethnics in the middle, and the lowest among “Hispanic”-identified youth. This latter result was most clearly seen among Latin American-origin youth in South Florida, where they constitute a numerical majority in dense ethnic enclaves that buffer them from external discrimination.

Still, it is important as well to underscore the fact that despite their growing awareness of racial and ethnic inequalities, most of the youth in the sample (almost two-thirds) continued to affirm a confident belief in the promise of equal opportunity through educational achievement. Even more tellingly, as Table 8 shows, 61% of these youths agreed in the 1992 survey that “there is no better country to live in than the United States,” and that endorsement grew to 71% three years later—despite a growing anti-immigrant mood in the country and especially in California during that period. Again there were significant differences by ethnic identity types, and even wider differences by national origin groups. Tellingly, the groups most likely to endorse that view were the children of political exiles who found a favorable context of reception in the U.S.: the Cubans and the Vietnamese. The groups least likely to agree with that statement were those who had most felt the weight of racial discrimination: the children of immigrants from Haiti, Jamaica and the West Indies. In reacting to their contexts of reception and learning how they are viewed and treated within them, the youths form and inform their own attitudes toward the society that receives them—and their own identities as well.

Race and Identity

Near the end of the follow-up survey, the respondents were asked to answer a semi-structured question about their “race.” They were given the option to check one of five categories: “white,” “black,” “Asian,” “multiracial,” or “other;” if the latter was checked, they had to write in their own words what that “other race” was. The results (also presented in Table 8) are revealing.

Less than half of the total sample checked the conventional categories of white, black, or Asian; 11% reported being multiracial; and over 40% checked “other.” When those “other” self-reports were coded, it turned out that nearly a quarter of the sample (24%) wrote down “Hispanic” or “Latino” as their “race,” and another 14% gave their nationality as their “race.” Among those youths adopting pan-ethnic identities, as Table 8 indicates, there is an obvious conflation of race and ethnicity in the way they define their identities: 92% of those who identify as “Asian” give that as their race, as do 85% of those who identify as “Black”—while 58% of those who identify as “Hispanic” or “Latino” extend racial meaning to that label as well. Among those youths identifying by the other ethno-national types, racial self-definition as “white” make a decisive difference: 58% of the plain-American youths said they were

white, compared to 17 % of the hyphenated Americans, and only 6% of those identifying by a foreign national origin.

The explicit racialization of the “Hispanic-Latino” category, as well as the substantial proportion of youths who conceived of their nationality of origin as a fixed racial category, are noteworthy both for their potential long-term implications in hardening group boundaries, and for their illustration of the arbitrariness of racial constructions. The latter point is made particularly salient by directly comparing the youths’ notions of their “race” with that reported by their own parents (the same “race” item was used in the parental interviews of 1995-96). Those results are cross-tabulated in Table 9, by national origin.

[Table 9 about here]

The closest match in racial self-perceptions between parents and children is evident among the Haitians, Jamaicans and other West Indians (most of whom self-report as black), among the Europeans and Canadians (most of whom label themselves white), and among most of the Asian-origin groups except for the Filipinos. The widest mismatches by far occur among the Latin-American-origin groups. For example, 93% of Cuban parents identify as white, compared to only 41% of their children; 68% of Nicaraguan parents see themselves as white, but only 19% of their children agree, as is the case among all other Latin-origin groups except Mexicans. The children of the former, instead, largely adopt “Hispanic” or “Latino” as a racial label, whereas scarcely any of their parents do so, as Table 9 vividly shows. Among the Mexicans, whose pattern differs from all of the others, the children preponderantly racialize the national label, whereas Mexican parents are more likely to use “other” (mestizo) and “multiracial” as descriptors.

These remarkable results point to the force of the acculturation process and its impact on children’s self-identities. Fully exposed to American culture and its racial definitions, children learn to see themselves more and more in these terms and even to racialize their national origins. We turn now to a multivariate analysis of these self-perceptions and their determinants.

Determinants of Ethnic Self-Identities

Thus far I have examined a number of variables that were theoretically and empirically linked to specific patterns of ethnic self-identification. Among these factors are parental socioeconomic resources, nativity and national origin, parent-child relations within the family, acculturation and language, discrimination and race. Which of these predictors have the strongest independent effects on the various ethnic and pan-ethnic identity outcomes? A series of multivariate logistic regressions (not shown) examined the effects of the same set of predictor variables on each identity type. To establish unambiguously the temporal ordering of effects, the predictors were measured in 1992, while the outcome variables were measured in 1995-96. Some of the main results are briefly summarized below.

The relationship of a few of these predictor variables to the identity outcome variables are first sketched in Figures 6 to 8. Figure 6 shows the clearly linear patterns of the relationship between socioeconomic status, acculturation, and discrimination in the continuum from a national-origin self-identity to hyphenated-American to plain American. Figure 7 sketches the relationship of the different identities to indices of the youths’ attitudes toward their parents and family. And Figure 8 depicts the manner in which identities are forged in a social field shaped by the interaction of two powerful social

forces, acculturation and discrimination, each pulling and pushing in different directions in the process of ethnic self-definition. The main types of ethnic and pan-ethnic identities have been mapped onto the space formed by the intersection of the two axes of acculturation and discrimination, based on their respective mean scores in the acculturation and discrimination indices. As Figure 8 shows, the national-origin identity occupies the high-discrimination, low-acculturation top left quadrant of the social field, while at the opposite end is found the plain-American identity, occupying the low-discrimination, high-acculturation bottom right quadrant. The hyphenated-American identity is located along the diagonal between those two, although closer to the American than to the national-origin location. In the low-discrimination, low-acculturation lower left quadrant is found the “Hispanic” pan-ethnic identity (and its variants). This identity was adopted preponderantly by youth in the Miami metropolitan area, where Latin American-origin groups form a majority of the population—and where an institutionally complete community has emerged that may serve both as a buffer against external discrimination and as a brake to rapid acculturation. Finally, in the high-discrimination, high-acculturation upper right quadrant are found second-generation youth who define themselves as “Black.” Their adoption of a pan-ethnic identity has little to do with a lack of acculturation—they clearly prefer English and American ways—but to persistent high levels of racial discrimination and the inexorable sense of otherness that accompanies it.

[Figures 6 to 8 here]

The patterns that emerge from the multivariate analyses confirm a continuum of determination from national-origin to hyphenated-American to plain-American self-definitions (the three main ethno-national identities), as well as distinct profiles for the three main pan-ethnic identities (Asian, Black, Hispanic). First, nativity variables remain powerfully linked to self-identity. Being born in the United States was the strongest predictor of identifying as a hyphenated-American, and a significant predictor of selecting a plain-American identity; it was also by far the strongest negative predictor of identifying by the parental national origin. Among pan-ethnic identities, being born in the U.S. was also predictive of identifying as Hispanic or Black (both of which are made-in-the-USA identities). Similarly, becoming a naturalized U.S. citizen positively predicts American self-definitions, with or without a hyphen, and negatively predicts a foreign nationality identity—but it is not associated with pan-ethnicity. Citizenship influences national self-definitions, over and above nativity, but it seems to be irrelevant for the adoption of racially-bounded minority identities. The acquisition of citizenship may be interpreted as signaling a stake in the society as a full-fledged member, legally as well as subjectively, with an accompanying shift in one's frame of reference. These two variables--nativity and citizenship--had much stronger effects on ethnic self-identification than our measure of years in the United States. Thus, it is not so much the length of residence in the receiving country but rather the nature of one's sociopolitical membership status in it that appears to be more determinative of the psychology of ethnic identity.

The parents' own nativity in turn exerts a strong independent influence on the ethnic socialization of the child. Having both parents born in the same nation strongly boosted the odds that the child will identify with the parents' nationality, or at least keep part of it by identifying binationally as a hyphenated-American. When the parents were not co-nationals, having one parent born in the United States also strongly increased the probability of adopting a hyphenated-American identity. Conversely, children whose parents were born in different (typically Latin American) foreign countries were more likely to identify pan-ethnically as Hispanic or Latino, showing again the impact of parental intermarriage on the lack of continuity of ancestral identities. As illustrated earlier in the case of Rosa,

the native-born daughter of a Cuban father and a Salvadoran mother, interethnic marriage blurs the boundaries of ethnicity for the children and clearly complicates and may derail the transmission of an ancestral identity.

Secondly, among the set of predictors involving the children’s perceptions of their parents and family, three such variables retained significant effects on ethnic identification: early perceptions of the parents’ own subjective ethnic identity, and the embarrassment and familism indices. Indeed, respondents who reported feeling embarrassed by--and hence not proud of--their parents were less likely to identify by national origin, and more likely to identify as a hyphenated-American. This index was also the strongest determinant of choosing an “Asian” self-identity: the more embarrassed of their parents, the more likely they were to adopt that pan-ethnic identity. By contrast, those high on the attitudinal familism index were more likely to identify by national origin. These results point to the varying effects on ethnic self-identity of the loosening of family ties—and of the maintenance of family attachments. In general, the above set of variables involving various aspects of the parent-child relationship underscores the fundamental influence of the family as a crucible of ethnic socialization in this diverse population of children of immigrants.

Thirdly, parental socioeconomic status washed out in most of these regressions, suggesting that its effect on modes of self-identification is indirect and more likely to be mediated through other variables. However, the two cases where direct effects are observable are noteworthy. Immigrant parents who are higher-status professionals were more likely to influence their children's selection of a foreign national identity. Conversely, a Hispanic minority identity was associated with lower-status parents, suggesting that, with other predictors held constant, more vulnerable low-SES youths were also more susceptible to that pan-ethnic mode of self-definition. In more resourceful, upper-middle-class immigrant families the child may have more reason to associate social honor with and to feel pride in the national identity of the parents—and to affirm it for themselves. Indeed, as Rosenberg has observed, the child's sense of self-worth is in part contingent on the prestige of the elements of social identity (1979: 13). This finding is in accord with theoretical expectations that identity shifts tend to be from lower to higher status groups and not vice-versa, all other things being equal.

Fourthly, acculturation and language use, preference and proficiency remain significantly but only moderately linked to most of the forms of ethnic self-identification. Youths who did not prefer English or American ways (low acculturation scores) and who were not proficient in English (as measured in the 1992 survey) were somewhat more likely to identify by a foreign national origin in the latest survey. Those who identified as Hispanic, by contrast, were youths who also did not prefer English or American ways (low acculturation) yet were proficient in English while also regularly speaking Spanish at home. Highly acculturated respondents who preferred and were proficient in English but who also used the foreign language frequently at home were most likely to identify binationally as hyphenated-Americans. If they were equally acculturated but spoke only English at home, they were more likely to drop the hyphen and identify as plain American. Yet the strongest effect on the probability of identifying as “Black” or “Black American” was the frequency of speaking only English at home (an identity selected primarily by native-English-speaking West Indians but not by Haitians whose mother tongue is not English). In sum, these identity choices reflect not only varying levels of linguistic fluency but also the nuances of how and which languages are used with significant others in varying social contexts. As noted earlier, the switch to English is part of a larger acculturation process bound to entail the abandonment not only of an ancestral language but also of an ancestral

identity. What the switch to English per se does not predict is the form that alternative self-identities will take.

Finally, discrimination and racialization also influence ethnic self-identification. Youths who scored high on the discrimination index were significantly less likely to identify as “American,” and more disposed to identify with a foreign national origin, without any “American” qualifier attached. Such experiences and perceptions of exclusion and rejection on ascribed racial-ethnic grounds clearly undercut the prospect of identificational assimilation into the mainstream. Moreover, when a “white” racial self-conception was entered into the equation, that variable had by far the strongest effect on the determination of plain American identities. An “American” ethnic identity and a “white” racial identity emerge as nearly synonymous in these results. What is more, when national-origin groups were entered in the models (as dummy variables), the only ones who retained strong and positive net effects on the odds of selecting a plain American identity were Europeans and Canadians. Without exception, none of the Latin American, Caribbean or Asian groups in the sample had a significant positive association with that particular identity outcome.

Among the various pan-ethnic identity types, high scores on the discrimination index were significantly associated only with the selection of a Black self-identity. But discrimination had a strong negative effect on the selection of a Hispanic identity, underscoring the observation made earlier about the comparatively protective aspects of the Miami enclave as a buffer against experiences and expectations of discrimination among Latin-origin youths. In this regard, a contextual variable—the type and location of the school (and by inference, of the neighborhood and the proximal peer groups) in which these youths were placed--made an additional difference in the process of self-definition. With all other predictor variables held constant, having attended inner-city schools in junior high school where most students were racial-ethnic minorities significantly increased the likelihood of developing a racial or pan-ethnic identity (particularly for youths reporting a Black or Hispanic-Latino self-identity), while decreasing the probability of identifying ancestrally by national origin. These results provide empirical support for a segmented assimilation perspective, here applied to the process of ethnic (and racial) self-definition.

Did any particular nationality continue to affect the types of ethnic self-identification once the other predictors were taken into account? In many cases they did not; that is, despite the wide range of differences between them, many nationalities (entered as dummy variables into the regression equations) retained no significant effects in the models tested. In some cases, however, they did—suggesting that there is more about those national-origin groups and their contexts of reception that is not explained by the set of predictors included in the model. For example, I already noted the strong independent effect of European and Canadian national origins on the selection of a plain American identity. More than any other nationality, Cubans stood out for their strong attachment to a hyphenated-American identity—a “life on the hyphen”(Pérez Firmat 1994)—and they were followed closely in this propensity by the Vietnamese, net of all other factors. Two other important cases involve the Mexicans and Filipinos, both of whom retained strong positive effects on the likelihood of adopting a foreign national identity.

As shown earlier (in Table 4), those two California-based nationalities shifted to the foreign national identity by a greater percentage than any other group on either coast. And in the case of the Mexicans in particular, anecdotal evidence was provided to illustrate how that the California sociopolitical context in which the divisive Proposition 187 campaign was conducted had heightened the salience of ethnic and racial boundaries and accompanying processes of reaction formation. As a test of

this effect, I entered into the regression a dummy variable for regional location. When this contextual factor was accounted for, the Mexican nationality effect completely washed out of the equation—but the Filipino effect, although reduced in magnitude, remained strong. The results suggest that the propensity of the Mexican–origin youths to identify by national origin at the time of the 1995 survey was accounted for by the effect of the California context. However, the similar propensity of Filipino youths was not explained by any of the predictor variables in the model, raising intriguing theoretical and empirical questions as to why this is so. Perhaps the history of U.S. colonialism in the Philippines and the valence of that colonial past in Filipino collective consciousness may have more to do with the identity choices of Filipino youth as forms of ethnicity that react as much against that past as to their present circumstances. Indeed, the larger theoretical significance of national-origin groups as proxies for their contexts of reception in the United States—and as proxies for the histories of the relations of those countries with the United States—further underscore the complexity of the factors that may impinge on the determination of ethnic self-identities in the second generation.

Expanding the CILS Study: Current Mixed-Methods Research

As noted earlier, with support from the Russell Sage Foundation, we returned to the field in 2001-03 to locate and contact as many of our respondents as possible (from the baseline sample drawn a decade ago in Fall 1991, when most were 14 or 15 years old)—something made much more feasible now with new technologies and internet companies that have vastly enhanced the success rate of computerized searches on the basis of a few key items of information. We succeeded in locating most of the original respondents, even though many of them (who were now 24 or 25 years old) were no longer living with their parents in the San Diego and Miami metropolitan areas and some, in fact, had moved to many different states across the country. The new (“T3”) survey focused on the social, cultural, and economic trajectories of this diverse sample of young adults over the previous six years since the last follow-up of 1995-96. It also included a checklist inventory of major experiences and life events (including contact with the criminal justice system) over the same time period. During these critical years, our respondents have been making their transitions to adulthood: completing their education, leaving the parental home and establishing independent households, securing stable employment, marrying or cohabiting, forming a family and becoming a parent, some becoming civically involved or even maintaining transnational ties.

While the results of this latest phase of the study are not yet available, preliminary findings show the wide range and ethnic segmentation of their fates to date: from dropouts to medical school graduates, from county jail to Ivy League schools, from aimless drifting in junior colleges and the labor market to well-paying full-time jobs in mainstream careers. The majority remain single and without children to date, but some already have as many as four or five children, whether married, cohabiting or single-parenting, and substantial proportions are involved in interracial and interethnic relationships. For the men, arrest and imprisonment are the life events most likely to derail them from the educational and occupational paths they had earlier aspired to, while having children early is especially likely to derail the women from their own youthful ambitions. They move back and forth rather than in a linear fashion from one status to the next, depending on a variety of factors, contexts, and turning points. We aim to map in detail the sequence and patterning of these trajectories in early adulthood, and analyze their general direction and determinants; trace the degree to which ethnic differences in achievement, acculturation, and experiences of discrimination in adolescence are maintained, widened or reduced in young adulthood; and probe their implications for social mobility, intermarriage, and ethnic self-identity.

As so many of our findings show, both subjectivity and situationality are essential to understanding the meaning and changes in ethnic self-definitions among these young new members of American society. Contexualization both helps us account for differences in survey data, and requires further study using non-survey methods such as depth interviews, oral histories, and ethnography. Accordingly, with support from the MacArthur Research Network on Transitions to Adulthood and Public Policy, we began a series of in-depth, open-ended, qualitative interviews with a stratified sample of up to 150 of our original San Diego respondents. Through these intensive interviews and oral histories—which are taped, transcribed, and coded via specialized computer software for qualitative analysis—we aim to explore and probe in much greater detail and nuance the kinds of issues that survey instruments and methods constrained us from delving into. Among the themes we have identified for detailed exploration is that of ethnic self-identity, including issues of situational ethnicity, change in ethnic self-identity over time and in different social contexts, the subjective meanings of ethnic and racial labels and identities to the respondents, and—with respect to the effect on ethnonational self-definitions of major historical events—their perceptions of and reactions to the terrorist attacks of September 11, 2001, in the United States, and the concerns which these have provoked about national membership, belonging and loyalty in a time of war. Given the time-consuming and labor intensive nature of these qualitative interviews, their transcription, coding and qualitative analysis will extend well into 2003, and then merged with the quantitative survey data that will by that time have been collected, coded, and processed.

Still, early results from these transcribed interviews are already providing a wealth of information about the dynamic processes and contexts that shape ethnicity. For instance, two Mexican-American men in their mid-twenties focused on very different facets of their modes of ethnic self-definition. One focused on the importance of asserting a Mexican identity to others: “It's important for me to say that I’m Mexican. I cross out, like when they give me something on an application to fill in what I am, and Mexican's not there, I put ‘other’ and I put ‘Mexican.’” For the other, poignantly, such an assertion was hard to maintain with his own family: “We were all pachucos… considered not Mexican because we were born in the U.S., and because we were Mexican and didn't know Spanish. We were looked down on by our own family members because we weren't born in Mexico and weren’t speaking Spanish.” A young man from Ecuador, by contrast, eagerly identified the U.S. as his homeland, a conviction cemented during a return trip when he was a senior in high school. He went with his family for a two-week vacation but “after three days, I was soooo ready to go home!” He hated the experience: lack of water for showers, limited electricity, no streets—“just dirt with a bunch of rocks.” Yet he strongly identifies as “Ecuadoran American,” he said, because most people in California think that he is Mexican or “Latino” and he wants them to know that he was born in Ecuador and his roots are there. He doesn’t want to forget where he came from.

Contexts shape the meanings of identity assignments and assertions. A young woman identified strongly as Mexican because it is “her blood roots.” Even though she went to a predominantly white high school, she said, she never wanted to be white. But she noted that when she went to Mexico, she did identify as American because her cousins imposed this ethnicity on her whereas when she is in the U.S., others tend to impose the Mexican identity on her, not “American.” A 25-year-old Vietnamese man put it this way: “I'm from Vietnam but I live in America… kind of a hybrid. I… identify more with one part or the other when I’m in the opposite setting, you know? So, if I'm in a very non-Vietnamese setting, I'll feel different and I'll identify more as Vietnamese. But when I was in Vietnam, I felt very American in some ways. When I felt tired of speaking only in Vietnamese, I could kind of hang out with these other foreign students and be kind of a tourist. And when you're a tourist in Vietnam, like in

a lot of other countries, then the cultural rules don't really apply. So it gave me a lot of freedom, ’cause I could kind of switch.”

Identities change and evolve over time and lived experience as well as social space. When he was growing up, a Vietnam-born young man told us that he considered himself as American, didn’t want to associate with other Vietnamese, and hated the Vietnamese Saturday language school that his parents sent him to. In high school most of his friends were “Anglos” and he had no Vietnamese friends. It was only when he attended the University of California that he was “exposed to a lot of cool Asians, Vietnamese, Chinese… and my friends were Asians.” This was when he started identifying himself as Vietnamese American and became more aware about his origins. A different case is Joseph, who currently identifies as “Asian American.” He chooses the label because it is a “broader category,” although this identity is not important to him at all. American to him means being part of the U.S. When he was younger, he identified as Chinese because “we spoke Cantonese at home.” When he found out that he was born in Vietnam, he identified as Vietnamese. When he grew up more, he identified as Chinese Vietnamese. Now, he identifies as Asian American because his primary language in English. And still another is the case of a young woman who while in high school considered herself Vietnamese as a way to explain her differences with the broader society. It was in college that she began to consider herself “Vietnamese-American: ethnically Vietnamese, culturally American,” adding that she has had the most conflict when dating Vietnamese males because she is expected to behave in accordance with female Vietnamese norms. In this manner, she has adopted the American ideal of gender equality over the “traditional foundation” provided by Vietnamese norms. If she has children, she says, she will raise them to identify as American.

The meaning of an asserted identity is not self-evident. For instance, a young woman told us that she identifies as “Filipino American.” Although she doesn’t know much about the culture, she is proud to be Filipino and to be living in the U.S. She prefers this identity to a pan-Asian label because she’d like others to appreciate individual cultural differences, not lump them all together. Before entering college, she used to identify only as “Filipino” because of “the pride thing,” she said, but since college, she includes “American” as part of her ethnicity “because ‘American’ means anyone living in the U.S., not just whites.” In sharp contrast, a young Cambodian woman insisted on defining “American” in racial terms, saying emphatically: “How could I be American? I black skin, black eyes, black hair… my English not good enough and my skin color black.” In further contrast, a young 26-year-old man identifies as “Filipino African American.” He notes that it is very important because “it says everything about me,” who he is. He is proud that he has two different ethnic backgrounds because it gives him a different outlook on how he sees life and other people. He identifies more as black because of his skin color, but was raised surrounded by Filipino food, language, and traditions, and remembers how his mother would compare everything here to how it was in the Philippines, a frame of reference he still applies. And for yet another respondent, who considers herself Chinese, the distinction between Chinese and Chinese-American depends on one’s birthplace as well as one’s knowledge of Chinese language and culture. While growing up in the U.S. (she came when she was eight) she only watched Chinese movies, until she discovered during her 1996 visit back home, much to her chagrin, that people in China listen to American music. Despite that “betrayal,” it was then that she began to accept American English. It is important that her significant other be able to communicate with her parents; therefore, she will only date and marry Chinese or a Chinese speaking Chinese-American. She notes that some of her friends only look Chinese, but they are Chinese-Americans who can’t talk to her parents.

As these brief examples illustrate, the study of ethnic identity is multi-layered and can gain a great deal from mixed methods in comparative research. By combining quantitative and qualitative approaches, and by extending our longitudinal study to encompass a full decade across a critical transitional period of the life span—we will significantly enhance and contextualize our understanding of identity formation processes. Although the final results of these new efforts will not become available for some time, one lesson we have learned through our research to date is plain: namely, the importance of employing mixed methods in order to be able to extend and deepen the analysis of complex issues such as those addressed in this chapter and in this book.

Conclusion

Ethnic identities are not inevitable outcomes, but complex products of people’s ongoing efforts to interpret, understand and respond to the social structural, cultural, and historical situations in which they find themselves, within their sets of resources and vulnerabilities. In the present era of civil rights, affirmative action and ethnic revivals—and in contrast to the way immigrants were treated during the heyday of hegemonic Americanization in the early 20th century—today’s newcomers are re-categorized in broad racialized pan-ethnic clusters that the host society deems appropriate for those sharing (or imagined to share) a particular language or phenotype—and that convey the symbolic message, with its attendant stereotypes, that they belong to a subordinate status in the national hierarchy. Although the state, the school system, the mass media, and the society at large insist on redefining these immigrants and their children into one-size-fits-all pan-ethnic labels such as “Hispanics” and “Asians,” the children themselves are quite plural in their ethnic self-definitions. Challenged to incorporate what is “out there” into what is “in here” and to crystallize a sense of who they are, they “translate themselves” and construct a variety of self-identities. Some cling tenaciously to their parents’ national loyalties and retain their parents’ national identities, with or without a hyphen. Others shift to an unhyphenated American self-image and identify symbolically with the mainstream, banking on their unaccented English and consonant phenotypic traits. Still others internalize the racial and pan-ethnic categories into which they are constantly classified and identify symbolically with national minorities.

As we have seen, the paths to those different forms of ethnic self-definition are shaped by a variety of social and psychological forces. The results of our surveys show major differences in patterns of self-identification among teenage children of immigrants from scores of sending countries growing up in two distinct corners of the United States. They suggest some of the complex, conflictive, often incongruous and unexpected ways in which race and class, discrimination and acculturation, family relationships and personal dreams can complicate their sense of who they are. And they suggest that identities are neither fixed nor irreversible, but always a function of relational processes, whose meaning is embedded in concrete social and historical contexts.

Indeed—to return to the four epigraphs with which we began this paper (by Taylor, Hoffman, Pérez Firmat, and Hwang)—in this dialogical process of “translating themselves,” some gain in translation, while others may be lost in it. In Life on the Hyphen, Gustavo Pérez Firmat wrote of the 1.5 generation of Cuban-Americans that “One-and-a-halfers are translation artists. Tradition bound but translation bent, they are sufficiently immersed in each culture to give both ends of the hyphen their due… One-and-a-halfers gain in translation” (1994:7). But—in the absence of perceived discrimination and reaction formation (which remain open empirical questions, as suggested by the unexpected responses reported among young Cuban Americans in Miami during the recent Elián González saga [cf. Santiago and Dorschner 2000])—the straight-line process of acculturation may well “thin” a Cuban

ethnicity and take its inevitable generational toll on his U.S.-born, Miami-raised children, if without interfering with their sanguine sense of belonging, as vividly portrayed in one of the epigraphs. Yet in another context, the experience of self-translation and the quandaries of identity can take on a “thickened” and quite different meaning, as expressed by Caroline Hwang, the daughter of Korean immigrants, whose reflections on race and marginality serve as a countervailing epigraph. If for some the search for identity may with time and acceptance blur and fade into the twilight of ethnicity, for others it may with sharpened salience lead into the high noon of ethnicity—and for still others something in the wide range between those poles. But the underlying process is one in which all children of immigrants are inescapably engaged and of which they are acutely aware: making sense of who they are and finding a meaningful place in their adoptive society.

Acknowledgements

I gratefully acknowledge the support provided by the William and Flora Hewlett Foundation for a fellowship year in residence at the Center for Advanced Study in the Behavioral Sciences, Stanford, during 2000-01; by the MacArthur Foundation Research Network on Transitions to Adulthood and Public Policy; and by research grants from the Russell Sage, Andrew W. Mellon, Spencer, and National Science Foundations to Alejandro Portes and Rubén G. Rumbaut, Principal Investigators, for the Children of Immigrants Longitudinal Study (CILS), 1991-2003.

REFERENCES Alba, Richard D. 1985. Italian Americans: Into the Twilight of Ethnicity. Englewood Cliffs, NJ: Prentice-Hall. Alba, Richard D. 1990. Ethnic Identity: The Transformation of White America. New Haven: Yale University Press. Aleinikoff, T. Alexander, and Rubén G. Rumbaut. 1998. “Terms of Belonging: Are Models of Membership Self-Fulfilling Prophecies?” Georgetown Immigration Law Journal, 13, 1: 1-24. Banks, Sandy. 1994. “Unflagging Controversy.” Los Angeles Times, November 10. Bozorgmehr, Mehdi. 1997. “Internal Ethnicity: Iranians in Los Angeles,” Sociological Perspectives, 40, 3: 387-408. Bozorgmehr, Mehdi, Georges Sabagh, and Ivan Light. 1996. “Los Angeles: Explosive Diversity.” In Silvia Pedraza and Rubén G. Rumbaut, eds., Origins and Destinies: Immigration, Race, and Ethnicity in America. Belmont, CA: Wadsworth. Cornell, Stephen, and Douglas Hartmann. 1998. Ethnicity and Race: Making Identities in a Changing World. Thousand Oaks, CA: Pine Forge.

Deaux, Kay. 1996. “Social Identification.” In E. Tory Higgins and Arie W. Kruglanski, eds., Social Psychology: Handbook of Basic Principles. New York: Guildford Press. Erikson, Erik H. 1968. Identity: Youth and Crisis. New York: W.W. Norton. Escobar, Gabriel. 1999. “Dominicans Face Assimilation in Black and White.” The Washington Post, May 14. Gans, Herbert J. 1979. “Symbolic Ethnicity: The Future of Ethnic Groups and Cultures in America.” Ethnic and Racial Studies 2, 1: 1-20. Gordon, Milton M. 1964. Assimilation in American Life: The Role of Race, Religion, and National Origins. New York: Oxford University Press. Handlin, Oscar. 1973 [1951]. The Uprooted. 2nd edition. Boston: Little, Brown. Hoffman, Eva. 1989. Lost in Translation: A Life in a New Language. New York: Penguin Books. Horowitz, Donald L. 1985. Ethnic Groups in Conflict. Berkeley: University of California Press. Hwang, Caroline. 1998. “The Good Daughter.” Newsweek, September 21. McDonnell, Patrick J. 1994. “Complex Family Ties Tangle Simple Premise of Prop. 187.” Los Angeles Times, November 20. Min, Pyong Gap. 1995. “Korean Americans.” In Pyong Gap Min, ed., Asian Americans: Contemporary Trends and Issues. Thousand Oaks, CA: Sage. Nahirny, Vladimir C., and Joshua A. Fishman. 1996 [1965]. “American Immigrant Groups: Ethnic Identification and the Problem of Generations.” In Werner Sollors, ed., Theories of Ethnicity: A Classical Reader. New York: New York University Press. Navarrette Jr., Ruben. 1994. “At the Birth of a New—and Younger—Latino Activism.” Los Angeles Times, November 13. Pérez Firmat, Gustavo. 1994. Life on the Hyphen: The Cuban-American Way. Austin: University of Texas Press. Phinney, Jean S. 1990. “Ethnic Identity in Adolescents and Adults: Review of Research.” Psychological Bulletin, 108, 3: 499-514. Portes, Alejandro, and Rubén G. Rumbaut. 1996. Immigrant America: A Portrait. 2nd edition. Los Angeles: University of California Press. Portes, Alejandro, and Rubén G. Rumbaut. 2001. Legacies: The Story of the Immigrant Second Generation. Los Angeles and New York: University of California Press and Russell Sage Foundation.

Pyle, Amy, and Simon Romero. 1994. “Prop 187 Fuels a New Campus Activism.” Los Angeles Times, October 25. Rosenberg, Morris. 1965. Society and the Adolescent Self-Image. Princeton, NJ: Princeton University Press. Rosenberg, Morris. 1979. Conceiving the Self. New York: Basic Books. Rumbaut, Rubén G. 1994. “The Crucible Within: Ethnic Identity, Self-Esteem, and Segmented Assimilation among Children of Immigrants.” International Migration Review 28, 4 (Winter): 748-794. Santiago, Fabiola, and John Dorschner. 2000. “Outside Image Bewilders Exiles.” The Miami Herald, April 23. Tajfel, Henri. 1981. Human Groups and Social Categories. London: Cambridge University Press. Taylor, Charles. 1992. “The Politics of Recognition.” In Amy Guttman, ed., Multiculturalism and ‘The Politics of Recognition’. Princeton, NJ: Princeton University Press. Tilly, Charles. 1996. “Citizenship, Identity, and Social History.” In Charles Tilly, ed., Citizenship, Identity, and Social History. Veerasarn, Oi. 1999. “The Faces of Asian America.” Common Quest, 4, 1: 46-55. Warner, W. Lloyd, and Leo Srole. 1945. The Social Systems of American Ethnic Groups. New Haven, CT: Yale University Press. Waters, Mary C. 1990. Ethnic Options: Choosing Identities in America. Berkeley: University of California Press.

Table 1 Size, Location (in Florida or California), and Family Socioeconomic Status

by National Origin of Respondents

National Origin

Total Sample

(N)

Miami/ Ft. Lauderdale

(N)

San Diego

(N)

Family Socioeconomic Status1

Latin America: Mexico 755 28 727 .269 Cuba (private school) 182 182 0 .612 Cuba (public school) 1,044 1,042 2 .455 Nicaragua 344 340 4 .443 Colombia Dominican Republic Other Latin America Haiti and West Indies:

227 104 366

223 100 311

4 4

55

.451

.416

.474

Haiti 178 177 1 .382 Jamaica, West Indies 272 253 19 .484 Asia:

Philippines 819 11 808 .516 Vietnam 370 8 362 .342 Laos (Lao) Laos (Hmong) Cambodia Chinese Other Asia

155 53 95 72

100

1 0 1

23 22

154 53 94 49 78

.233

.103

.172

.501

.551

All Other Countries2 126 120 6 .559

TOTAL 5,262 2,842 2,420 .424

1 Composite index (0 to 1) of father’s and mother’s education, occupational prestige, and home ownership. 2 Of those 126 respondents, 88 are of European or Canadian origin, while 38 hail from the Middle East or Africa. Source: Children of Immigrants Longitudinal Study.

Table 2 Family Structure and Perceptions of Family Relationships, 1995-96, by National Origin and Children’s Acculturation and Language Ability

Characteristics Intact

Family1 (%)

Family Cohesion2

(% high)

Parent-Child Conflict3

(% high)

Embarrassed of Parents4

(%)

Familism5 (% high)

National Origin: Cuba (private school) 69.5 41.1 35.6 26.7 24.0 Cuba (public school) 55.1 35.8 38.3 28.9 24.3 Dominican Republic 46.1 38.5 35.9 14.1 24.4 Mexico 55.4 38.6 32.2 14.2 31.2 Nicaragua 58.4 42.0 31.7 21.7 25.6 Colombia 57.3 37.3 35.1 23.2 28.6 Other Latin America 59.3 40.4 33.9 28.9 24.3 Haiti 42.2 22.2 57.0 41.5 31.1 Jamaica 39.0 28.8 39.8 22.0 24.6 Other West Indies 42.2 30.1 41.0 38.6 20.5 Vietnam 71.3 27.7 49.4 38.1 43.9 Laos (Lao) 70.8 32.6 42.4 30.6 50.0 Laos (Hmong) 56.0 30.0 66.0 48.0 46.0 Cambodia 60.7 24.7 55.1 43.8 33.7 Philippines 76.9 28.6 46.7 33.7 24.3 Chinese, other Asia 77.4 34.2 37.4 52.3 21.9 Europe, Canada 59.6 28.1 26.3 35.1 21.0

p* < .001 < .001 < .001 < .001 < .001 Acculturation Index: 6 Low (< 0.33) 60.4 42.2 34.5 18.1 37.3 Middle (0.50) 61.4 35.7 37.2 24.7 28.3 High (> 0.66) 61.9 29.6 43.9 36.9 24.4

p* N.S. < .001 < .001 < .001 < .001 Language Dominance: 7 Fluent bilingual 58.8 47.0 30.6 21.3 23.0 English dominant 64.5 27.9 43.1 36.1 23.5 Foreign lang. dominant 54.2 42.7 33.3 17.4 42.1 Limited bilingual 61.9 26.3 48.6 31.3 37.8

p* < .001 < .001 < .001 < .001 < .001 TOTAL 61.4 34.0 39.9 29.3 28.2

1 Both biological parents present in the home at both surveys. 2 High score (> 4.0) in 3-item family cohesion index; items measured from 1 to 5. 3 High score (> 2.0) in 4-item parent-child conflict index; items measured from 1 to 4. 4 Embarrassed or ambivalent over parents’ ways. 5 High score (> 2.0) in 3-item familial obligations index; items measured from 1 to 4. 6 Score on composite index of child’s preferences for English language and American ways reported at both surveys. 7 Based on scores in English proficiency and foreign language proficiency indices, 1995-96. * Probability of differences being due to chance. N.S. = not significant. Source: CILS (longitudinal sample, N=4,288).

Table 3 Continuity and Change in the Ethnic Self-Identities

of Children of Immigrants, 1992 and 1995-96*

Ethnic Self-Identity in 1995-96

Total

Ethnic Self-

Identity Reported in 1992

National

origin

Panethnic

Hyphen-ated

American

American

Mixed, other

N in 1992

(%)

National origin

57.7

20.2

17.4

1.2

3.6

1,181 (27.5)

Panethnic

32.5

46.5

16.8

0.6

3.5

677

(15.8) Hyphenated American

28.1

20.6

44.6

2.7

4.0

1,748 (40.8)

American

10.7

33.8

32.8

14.8

7.9

542

(12.6) Mixed, other

22.9

27.9

26.4

3.6

19.3

140 (3.3)

Total

N in 1995-96 (%)

1,482 (34.6)

1,135 (26.5)

1,314 (30.6)

150 (3.5)

207 (4.8)

4,288 (100.0)

* Figures are row percentages, indicating the proportion of respondents who reported in the 1995-96 survey the same or a different type of ethnic identity as that given in 1992. Overall, 44% of the respondents reported the same identity in the follow-up survey, while 56% reported a different type. See text for a description of ethnic identity types. Source: CILS (longitudinal sample, N=4,288).

Table 4

Ethnic Self-Identities of Children of Immigrants, by National Origin Groups, 1995-96

Type of Ethnic Self-Identity*

National Origin Groups

National %

in 1995-96

Origin (% change since 1992)

Panethnic %

in 1995-96

Identities (% change since 1992)

Hyphenated %

in 1995-96

-American (% change since 1992)

American %

in 1995-96

Identity (% change since 1992)

Latin America: Cuba (private school) 6.2 (+ 2.7) 19.2 (+15.1) 70.5 (+13.0) 2.1 (-32.9) Cuba (public school) 16.5 (- 0.4) 31.5 (+23.2) 42.0 (-11.1) 6.1 (-14.6) Dominican Republic 5.1 (-20.5) 71.8 (+44.9) 10.3 (-15.4) 2.6 (-16.7) Mexico 41.2 (+23.5) 25.0 (-21.0) 28.9 (- 2.0) 1.2 (- 1.8) Nicaragua 54.1 (+19.6) 25.6 (-13.2) 17.4 (+ 2.1) 0.4 (- 8.2) Colombia 15.7 (- 7.6) 62.7 (+36.8) 13.0 (-17.3) 3.2 (-15.7) Other Latin America

16.7 (-10.0) 43.3 (+33.2) 23.3 (- 7.1) 13.3 (-16.7)

Haiti and West Indies: Haiti 37.8 (+ 6.7) 4.4 (- 7.4) 43.7 (+ 5.2) 0 (-14.1) Jamaica 39.0 (-13.6) 15.3 (+10.2) 29.7 (+ 2.5) 2.5 (-10.2) Trinidad, Other W.I.

14.5 (-13.3) 55.4 (+47.0) 16.9 (-10.8) 2.4 (-24.1)

Asia: Vietnam (Vietnamese) 57.9 (+13.1) 11.9 (+11.5) 29.4 (-18.3) 0 (- 4.0) Vietnam (ethnic Chinese) 25.9 (- 1.7) 29.3 (+29.3) 39.7 (-10.3) 1.7 (- 1.7) Laos (Lao) 66.7 (+ 4.9) 11.8 (+ 9.0) 19.4 (- 9.7) 0.7 (+ 0.0) Laos (Hmong) 48.0 (-14.0) 38.0 (+36.0) 12.0 (-14.0) 0 (- 4.0) Cambodia 48.3 (+ 7.9) 21.3 (+19.1) 30.3 (-15.7) 0 (- 3.4) Philippines 55.1 (+24.0) 1.9 (+ 1.2) 37.0 (-22.8) 1.7 (- 3.6) Chinese, Other Asia

19.4 (-12.9) 43.9 (+42.6) 23.2 (-24.5) 5.8 (- 8.4)

Europe, Canada: 8.8 (-14.0) 7.0 (- 3.5) 10.5 (-10.5) 57.9 (+15.8)

Totals

34.6 (+ 7.0) 26.5 (+10.7) 30.6 (-10.1) 3.5 (- 9.2)

* See text for description of ethnic self-identity types. Not shown is a residual category of mixed identities chosen by 4.8% of the respondents in the 1995-96 survey. Figures are row percentages.

Source: CILS (longitudinal sample, N=4,288).

Table 5 Nativity Patterns of Children of Immigrants and of their Parents

Nativity of Children* Nativity of Father and

Mother*

National Origin Groups

Foreign-born (1.5 generation)

%

U.S.-born (2nd generation)

%

Both born in same country

%

In different countries

%

One parent born in U.S.

% Latin America:

Cuba (private school) 8.9 91.1 85.6 9.6 4.8 Cuba (public school) 32.2 67.8 74.8 14.1 11.1 Dominican Republic 32.1 67.9 79.5 1.3 19.2 Mexico 38.2 61.8 73.0 9.5 17.5 Nicaragua 92.9 7.1 85.8 13.2 1.1 Colombia 48.1 51.9 64.9 25.9 9.2 Other Latin America 53.9 46.1 74.3 7.1 18.6 Haiti and West Indies:

Haiti 54.8 45.2 85.9 9.6 4.4 Jamaica 63.6 36.4 78.0 11.9 10.2 Trinidad/other 32.5 67.5 50.6 24.1 25.3 Asia:

Vietnamese 84.2 15.8 89.0 8.1 2.9 Laos (Lao) 98.6 1.4 95.1 4.9 0 Laos (Hmong) 94.0 6.0 90.0 10.0 0 Cambodia 96.6 3.4 80.9 19.1 0 Philippines 42.5 57.5 79.1 3.9 17.0 Other Asia 43.9 56.1 64.5 7.1 28.4 Europe, Canada: 14.0 86.0 22.8 3.5 73.7

Totals

49.9 50.1 76.8 10.4 12.9

* Figures are row percentages. Source: CILS (longitudinal sample, N=4,288).

Table 6 Nativity, Social Status, Language, Acculturation, and Attitudes toward Parents: Selected

1992 Predictors of Ethnic Self-Identities Reported by Children of Immigrants in 1995-96

Ethnic Self-Identity in 1995-961

Predictor Variables, 19922

National Origin

%

Panethnic

%

Hyphenated American

%

American

%

TOTAL SAMPLE

% Child’s nativity:

Foreign-born 74.5 46.6 31.1 8.0 49.9 U.S.-born 25.5 53.4 68.9 92.0 50.1

Parents’ nativity: Both born in same country 88.9 72.3 75.7 42.0 76.7 Born in different countries 6.5 15.8 7.5 16.0 10.4

One parent born in U.S. 4.6 11.9 16.8 42.0 12.9 If both parents foreign-born, year of immigration to U.S.:

Before 1970 4.5 14.2 20.9 40.2 13.0 In 1970s 15.8 26.8 35.4 37.9 25.8 In 1980s 79.7 59.0 43.7 21.8 61.2

Socioeconomic status:

Family owns home 48.7 52.8 66.4 73.3 56.6 Parent is a professional 20.6 24.3 26.2 40.0 24.3

Education 12yrs+ (mother) 53.3 58.9 64.6 76.7 59.7

Language and acculturation: Speaks foreign lang. “very well” 37.9 38.5 26.9 14.0 33.0

Speaks English “very well” 66.6 79.1 84.6 95.3 77.4 Parent speaks English “very well” 22.3 26.9 38.5 59.4 30.3

Speaks English with parents 32.7 32.0 48.7 70.0 40.2 Speaks English with friends 30.3 27.7 34.4 62.0 32.8

Prefers English 64.4 71.9 79.8 91.3 72.9

Prefers “American ways” 34.8 39.9 46.4 74.7 41.9 Parents prefer “American ways” 19.8 23.5 28.1 60.1 25.6

Attitudes toward parents/family:

Embarrassed of parents, 1992 22.9 23.4 27.1 30.0 24.6 Embarrassed of parents, 1995-96 26.5 28.3 31.2 42.0 29.3

Believe that only relatives can help with serious problems, 1992

19.6

16.4

12.2

7.3

15.5

Believe that living close to family is “very important”, 1995-96

46.8 42.7 39.7 24.8 42.0

1 Figures are column percentages. The probability that the differences by type of identity are due to chance is less than 1 in 1,000 for all variables in the table. The ethnic self-identities are those reported in the follow-up survey of 1995-96. For simplicity of presentation, a small residual category of mixed identities is not included. 2 Predictor variables were measured in the baseline survey of 1992, unless otherwise noted; parent’s English ability was measured at the time of the separate parental interviews in 1995-96. Source: CILS (longitudinal sample, N=4,288).

Table 7

Patterns of Ethnic Socialization within the Family: Correlations of Children’s Perceptions of Their Parents’ Ethnic Identities in 1992,

with Their Own Ethnic Self-Identities in 1992 and 1995-96

Parents’ Identity as Perceived by Child in 1992

Children’s

Ethnic Self-Identities

Time of survey

National origin

Hyphenated American

American

Panethnic

Mixed identity

National origin

1992:

1995-96:

.373*** .206***

-.229*** -.085***

-.202*** -.167***

-.058***

-.032*

-.024 .020

Hyphenated American

1992:

1995-96:

-.192*** -.120***

.280*** .131***

.005

.059***

-.066*** -.043**

-.045**

-.053*** American

1992:

1995-96:

-.201*** -.089***

.015 -.002

.251*** .128***

-.028

-.031*

-.023 -.016

Panethnic

1992:

1995-96:

.015 -.029

-.088***

-.034*

-.019 .030*

.186*** .083***

.050**

.022 Mixed identity

1992:

1995-96:

-.063*** -.063***

-.045**

-.021

.063*** .072***

.002 .020

.123*** .039**

*** Probability of correlation being due to chance is less than 1 in 1,000. ** Probability of correlation being due to chance is less than 1 in 100. * Probability of correlation being due to chance is less than 5 in 100. Source: CILS (longitudinal sample, N=4,288).

Table 8

Location, Experiences of Discrimination, Perceptions of the United States, and Self-Reported Race, by Main Types of Ethnic Self-Identities Reported by Children of Immigrants in 1995-961

Ethnonational Self-Identity in 1995-96

Panethnic Self-Identity in 1995-96

Predictor Variables

National Origin

%

Hyphenated American

%

American

%

Hispanic, Latino2

%

Asian

%

Black

%

TOTAL SAMPLE

% Location:

Southern California 67.3 47.6 21.3 19.8 75.1 16.3 48.1 South Florida 32.7 52.4 78.7 80.2 24.9 83.7 51.9

School attended in 1992:

Inner-city school 37.4 30.4 26.7 38.4 43.0 48.8 35.4 Suburban school 62.6 69.6 73.3 61.6 57.0 51.2 64.6

Experienced discrimination:

Reported in 1992 60.3 53.5 44.0 44.7 57.0 68.8 54.5 Reported in 1995-96 66.4 62.0 52.7 51.4 66.8 80.0 62.2

Expects discrimination:

Reported in 1992 36.6 33.0 22.7 21.7 33.2 61.3 32.0 Reported in 1995-96 39.7 35.3 27.3 25.4 34.2 62.5 35.1

Agrees that “The U. S. is the best country to live in:”

Reported in 1992 57.7 65.9 73.3 58.3 66.8 47.5 60.8 Reported in 1995-96 67.9 76.0 78.0 72.2 72.0 54.4 71.3

Child’s self-reported “race:”

White 6.0 17.4 58.7 20.8 - - 14.3 Black 5.7 6.7 2.7 0.8 - 85.0 6.6 Asian 40.4 23.1 3.3 - 92.2 - 25.8

Multiracial 5.1 15.0 16.0 11.5 6.2 13.8 11.4 Hispanic, Latino 16.4 17.2 10.0 57.8 - - 23.5

Nationality 23.3 18.0 2.7 4.8 0.5 - 14.9 Other response 3.2 2.7 6.7 4.4 1.0 1.3 3.6

Parent’s self-reported “race:”

White 16.8 32.9 63.8 60.8 0.9 4.7 30.2 Black 6.2 7.4 1.4 1.4 0.9 81.4 7.4 Asian 39.7 29.9 13.0 0.7 76.9 2.3 29.2

Multiracial 8.9 8.6 10.1 15.6 0.9 2.3 9.6 Hispanic, Latino 2.6 2.4 1.4 6.7 - - 3.2

Nationality 20.1 13.0 2.9 6.3 18.8 - 14.2 Other response 5.7 5.9 7.2 8.4 1.7 9.3 6.3

1 Figures are column percentages. The probability that the differences by type of identity are due to chance is less than 1 in 1,000 for all variables in the table. The types of self-identities are those reported in the 1995-96 survey. For simplicity of presentation, a small residual category of mixed identities is not included. 2 Reported self-identity as “Hispanic,” “Latino,” or “Chicano.” Source: CILS (longitudinal sample, N=4,288).

Table 9 Self-Reported “Race” of Children of Immigrants and their Parents, by National Origin Groups, 1995-96

Self-Reported Race”*

National Origin Groups

Respondent (Child/Parent)

White (%)

Black (%)

Asian (%)

Multiracial (%)

Hispanic, Latino

(%)

Nationality (%)

Other (%)

Latin America: Cuba Child 41.2 0.8 - 11.5 36.0 5.5 4.9 Parent 93.1 1.1 0.3 2.5 1.1 0.5 1.4 Mexico Child 1.5 0.3 - 12.0 25.5 56.2 4.5 Parent 5.7 - 2.1 21.6 15.9 26.1 28.5 Nicaragua Child 19.4 - - 9.7 61.8 2.7 6.5 Parent 67.7 0.5 1.6 22.0 5.4 0.5 2.2 Other Latin America Child 22.8 1.9 - 14.7 52.9 4.6 3.1 Parent 69.5 4.6 0.8 17.8 2.3 1.9 3.1 Haiti and West Indies: Haiti Child - 75.9 - 8.4 - 9.6 6.0 Parent - 85.5 1.2 - - 6.0 7.2 Jamaica, West Indies Child 3.4 66.4 7.6 15.1 - - - Parent 8.4 65.5 5.0 8.4 - 6.7 5.9 Asia: Philippines Child 1.1 - 61.6 13.2 - 23.0 1.1 Parent 0.3 0.5 44.1 11.1 - 41.4 2.7 Vietnam Child - - 89.8 1.6 - 7.0 1.6 Parent - - 99.6 - - - 0.4 Laos, Cambodia Child 0.4 - 87.8 3.4 - 7.2 1.1 Parent - - 74.9 - - 23.6 1.5 Other Asia Child 1.5 - 82.4 13.2 - 2.9 - Parent 2.9 1.5 76.5 1.5 - 8.8 8.8 Europe, Canada: Child 76.0 - - 8.0 12.0 - 4.0 Parent 84.0 - - 8.0 4.0 - 4.0

Totals Child 12.1 6.5 32.1 10.3 20.3 15.5 3.2 Parent

30.2 7.4 29.2 9.6 3.2 14.2 6.3

* Figures are row percentages. Source: CILS.

Figure 1 Family Indices and Level of Acculturation, 1995-96*

0

5

10

15

20

25

30

35

40

45

Family Cohesion (high) Parent-Child Conflict(high)

Embarrassed of Parents Familism ( high)

LowMiddleHigh

Acculturation:PERCENT

Family Indices, 1995-96

* Probability of differences being due to chance is less than 1 in 1,000 for all associations shown. Level of acculturation is measured from a composite index of child’s preferences for English language and American ways reported at both surveys. See text for details of family indices.

Figure 2

Ethnic Self-Identity Shifts among Children of Immigrants, 1992 to 1995-96

27.5

15.8

40.8

12.6

34.6

26.5

30.6

3.5

0

5

10

15

20

25

30

35

40

45

National Origin Panethnic Hyphenated American American

19921995-96

PERCENT

Ethnic Self-Identity Types*

* See text for description of ethnic self-identity types. Not shown are mixed identities chosen by about 4% of the respondents in both surveys. Source: CILS (longitudinal sample, N=4,288).

Figure 3 Percent of Children of Immigrants Selecting Various Types

of Panethnic Identities, 1992 and 1995-96

0.90.3

12

0.5

2.21.9

4.5

17.9

1.40.8

0

2

4

6

8

10

12

14

16

18

Black Asian Hispanic Latino Chicano

19921995-96

PERCENT

Type of Panethnic Identity*

* See text for description of panethnic self-identity types. Source: CILS (longitudinal sample, N=4,288).

Figure 4 Stability and Salience of Ethnic Self-Identities, 1995-96

57.7

46.5 44.6

14.8

71.4

52.857.2

42

0

20

40

60

80

National Origin Panethnic HyphenatedAmerican

American

Stability: percent reporting same identity 3 to 4 years laterSalience: percent reporting ethnic identity as "very important"

PERCENT

Ethnic Self-Identity Types

Source: CILS (longitudinal sample, N=4,288).

Figure 5

Ethnic Self-Identities, 1995-96, by Native Origin of Self and Parents*

0

10

20

30

40

50

60

70

80

90

100

Foreign-born U.S.-born Both parents born insame country

One parent born inU.S.

National OriginPanethnicHyphenated AmericanAmerican

Ethnic Identities:PERCENT

Native Origin of Self and Parents

* Probability of differences being due to chance is less than 1 in 1,000 for all associations shown.

Figure 6 Socioeconomic Status, Acculturation, and Discrimination, by Ethnic Self-Identities, 1995-96

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

Socioeconomic statusindex*

Acculturation index** Discrimination index***

National OriginPanethnicHyphenated AmericanAmerican

INDEX SCORE

* Standardized co posite index (0 to 1) of parents’ education, occupational status, and home ownership m

** Composite index (0 to 1) of preferences for English language and American ways reported at both surveys *** Composite index (0 to 1) of experiences and expectations of discrimination reported at both surveys

Ethnic Self-Identities:

Probability of differences being due to chance is less than 1 in 1,000 for all indices.

Figure 7 Indices of Family Attachment: Attitudinal Familism, Embarrassed of Parents,

and Desire to Live Close to Parents, by Ethnic Self-Identities, 1995-96

0

0.1

0.2

0.3

0.4

0.5

Familism index* Embarrassed of parents** Desire to live close toparents***

National OriginPanethnicHyphenated AmericanAmerican

Ethnic Self-Identities:

INDEX SCORE

* Composite index (0 to 1) of attitudes reported at both surveys. See text for description of items. ** Composite index (0 to 1) of attitudes reported at both surveys. See text for description of items. *** Proportion who rank living close to parents as “very important,” 1995-96

Probability of differences being due to chance is less than 1 in 1,000 for all indices.

Figure 8 Patterns of Ethnic Self-Identification among Children of Immigrants,

by Levels of Acculturation and Discrimination

Acculturation Index* LOW HIGH HIGH

Discrimination Index** Hispanic

LOW

Asian

Immigrant Nationality

Black

Hyphenated American

American Hispanic

* Acculturation index = composite measure (0 to 1) of preferences for English language and American ways reported at both surveys. ** Discrimination index = composite measure (0 to 1) of experiences and expectations of discrimination reported at both surveys. See text for description of ethnic self-identity types, and the appendix for mean scores in these two indices by national origin.